A novel diagnostic technique named “Tracer Producing LIF technique” which permits 2-dimensional measurements of internal EGR in an engine cylinder has been developed. The feature of this technique is utilizing a fuel additive which does not emit LIF signal but its combustion products radiate strong LIF emission. Internal EGR behaviors can be measured by observing the LIF images excited by a UV-laser sheet. Firstly, the principle of this technique and selection of fuel additives are described. Then, "Tracer Producing LIF technique" was applied to the visualized engine, in which whole pent-roof area can be observed from the engine side direction. Internal EGR behaviors have been measured from suction to exhaust in whole engine cycle. As a result, gas exchange process between burned gas and air, as well as the influence of the valve timing on the above process were clarified. It was also confirmed that inhomogeneous internal EGR distribution still existed at the ignition timing for the first time.

Cycle-resolved measurements of the fuel concentration near a spark plug in a practical spark-ignition (SI) engine have been developed. An in situ laser infrared (IR) absorption method was applied using a spark plug sensor and a 3.392-μm He-Ne laser as the light source. The newly developed IR spark plug sensor had a higher signal-to-noise ratio than its previous version due to the optimization of its sapphire lens and two optical fibers. Firstly, a database of the molar absorption coefficients of regular and premium gasoline at different pressures and temperatures was established in advance using constant-volume vessel. The molar absorption coefficient of gasoline decreased with increasing pressure above atmospheric pressure. Secondary, the measurement accuracy was investigated in a constant-volume vessel and a compression-expansion machine. Thirdly, fuel concentration measurements in a spark-ignition (SI) engine with ethanol blended gasoline were carried out using an optical sensor installed in the spark plug with laser infrared absorption technique. Mixture formation processes of ethanol blended gasoline were investigated using spark-plug sensor installed in a spark plug in a port-injected SI engine with changing the fuel injection timing in an intake port.

The aim of this paper is the understanding of knocking occurrence mechanism from the view of flame propagation behavior by using 3-D simulation with UCFM. The flame propagation behavior impacted by in-cylinder flow was analyzed by calculated results and experimental visualizations. The flows of tumble and swirl motion were given to in-cylinder by setting various baffle plates in the middle of intake port. The comparisons of the measured and calculated flame propagation behavior show good agreement for various conditions. The results indicate that in-cylinder flow varies flame propagation shape from initial combustion, and it has strong influence on knocking occurrence. In addition, the influence of chemical reaction in end-gas which is generated by the in-cylinder flow was analyzed.

General overview of the combustion models for a diesel engine is described. The details of auto-ignition and mixing controlled combustion models are explained, especially focused on the available ranges of these models from the viewpoint of the ratio between chemical reaction rate and turbulent mixing rate and non-homogeneity of the mixture. In addition, a diesel auto-ignition model is explained, which is based on the auto-ignition process of non-homogeneous mixture.

To relate actual turbulent premixed flames observed in many applications with turbulent combustion diagram, experimental method for evaluations of turbulence characteristic length and Reynolds number based on Taylor microscale are discussed in detail. As for the hot wire anemometer, importance of energy spectrum and probability density function of velocity derivatives are shown for the accurate prediction of Taylor microscale. In the case of particle image velocimetry (PIV), spatial resolution should be in the same order of those of direct numerical simulation (DNS) of turbulence (less than 3-4 times the Kolmogorov length scale) to obtain Taylor microscale accurately. Current state and perspective of DNS of turbulence and turbulent combustion are discussed with feature trend of the fastest supercomputer in the world. Based on the perspective of DNS of turbulent combustion, possibility of perfect simulations of IC engine is shown. In 2020, the perfect simulation will be realized with 30 billion grid points by 1EXAFlops supercomputer, which requires 4 months CPU time. The CPU time would be reduced to about 4 days if several developments are achieved in the current fundamental researches.

This review paper, which is based on the survey about bio fuels in the committee of the Combustion Society of Japan (2005-2007) and in addition with our recent experimental researches, presents the potential of advanced technologies of bio fuel engines.

It is shown that the ethanol fueled chemical turbo engine (CTE) has good promise of performance and efficiency gains. The CTE essentially has two advantages over usual spark ignition engines. First advantage is recover of heat of exhaust gas as chemical energy of gas mixture can be produced by reforming of ethanol at comparatively low temperature, around 600K. With the advances in reformer technology, the use of higher fractions of hydrogen and other flammable species, could provide the second advantage of CTE i.e. Increase of the indicate work due to the higher mass burning rate.

We also experimentally evaluate a performance and exhaust emissions of diesel engine without any modifications operating on BDF (biodiesel fuel) reforming by ozone-rich air generated by corona discharge ozonizer. The combustion characteristics, exhaust emissions are compared with the case of BDF as baseline fuel. It is expected that the flammable ozonide is generated in double bond of fatty acid contained in BDF by fuel reforming. On the other hand, the ozone rich air is mixing into intake air of diesel engine as oxidizer, and exhaust emissions of engine fueled with BDF are studied.

This paper proposes a unified model that includes an effect of preferential diffusion of hydrogen. As hydrogen has a larger diffusion coefficient than methane and oxygen, the mass fraction of hydrogen varies in a premixed flame which is with curvature. This phenomenon is well known as preferential diffusion of hydrogen. Kido and Nakahara reported that the preferential diffusion influences the blow off limit of turbulent premixed flame and estimated the empirical formula due to experimental data. To simulate combustion phenomena of premixed gas that includes hydrogen with high accuracy, the preferential diffusion should be considered on the combustion model. On this paper, a simple model of the diffusion process of fuel in the premixed gas was evaluated firstly. Next, the results were applied to the author's combustion model. The results were compared with the experimental data by Kido and Nakahara and were in good agreement.

Soot is a very small particle formed near a combustion field, where a very steep temperature gradient exists. It is important for appropriate control of combustion devices to understand the behavior of soot particles in combustion fields. In this study, the behavior of each soot particle under temperature gradient is observed in order to reveal the thermophoretic behavior of soot particle. The relationship between the morphological character and the measured thermophoretic velocities of soot particles was examined quantitatively. Then, the method to estimate the behavior of soot particle induced by thermophoresis was examined. The dimensionless density, the ratio of the bulk density to the true density, was suggested as one of the morphological character of soot particle. It was found that the relationship between the thermophoretic coefficient, K_th, and the dimensionless density is a first-order negative correlation especially when the size of primary particle is less than 50nm. It was found that the thermophoretic velocity becomes slower as the dimensionless density becomes greater, that is, the aggregated particles are of a closed structure. The thermophoretic coefficient approached an asymptotic value, K_th = 0.53, which is almost same as the value for the particles in free molecule regime, as the dimensionless density becomes smaller. These results enable to estimate the thermophoretic velocity from the dimensionless density of the aggregated particles. This will be useful for the appropriate estimation and control of soot particle near a combustion field and the development of combustion devices.

Catalytic combustion method was studied to remove CH₄ efficiently in the off-gas from fuel reforming type FC system. The higher temperature was necessary to remove CH₄ by using the combustion catalysts so that the off-gas contained relatively high degree of moisture. The effective removal of CH₄ under the condition of lower temperature was possible according to the O₂ control method which adjusted O₂ amount to lower than the stoichiometry of CH₄ oxidation. In this case, H₂ and CO were detected in the catalyst outlet gas, so it has been understood that the partial oxidation reaction has occured. The catalyst surface was thought to adsorb O₂ easily than CH₄, and the O₂ covered the catalyst surface, therefore the CH₄ combustion was inhibited when the amount of O₂ increased. The O₂ control method was thought to be effective to exclude the inhibition. Coexisting H₂ in the off-gas raised the catalyst bed temperature by the H₂ oxidation so that it promoted the CH₄ conversion. The influence of CH₄ concentration was also discussed. The comparison was made with data for Pd and Pt catalyst performance. The effects of loading amount of Pd on the catalyst performance were discussed, and it was suggested that the H₂ and CO oxidation, CO shift reaction and steam reforming could be occurred following the partial oxidation of CH₄. So it is important to promote the series of these reactions for the efficient removal of residual CH₄ in the wasted gas.

Pulse detonation engines (PDEs) are expected to be the next-generation engine systems, and are expected to have applications in various fields. One of the fundamental problems faced during the development of PDEs is the deflagration to detonation transition (DDT). In order to realize a PDE, it is essential to shorten a parameter of detonation transition called a detonation induction distance (DID). However, the mechanism underlying the DDT process and methods to shorten DID in narrow channels have not yet been understood. An experimental study on DDT process in a narrow rectangular channel with a height of 1-5 mm and a width of 8 mm was carried out by employing pressure transducers, ionization probes, and sooted plate technique in oxyhydrogen mixtures. The effects of the tube height and equivalence ratio on pressure detonation limits were discussed. Detonation velocity, DDT process, and DID were discussed on the basis of pressure history and soot track record. Overdriven detonation and attenuated detonation were observed in the narrow channel. The DID value measured on the basis of the soot track record in the present study was in the range of the empirical formula obtained by the other researchers.