The objective of this study is to develop a dynamic subgrid scale model, for which computational results will not depend on the grid resolution. The most commonly used SGS model is based on the Smagorinsky eddy viscosity model with the model coefficient computed dynamically through the tensor level identity by Germano et al. However, the tensor level identity does not explicitly include the effect of grid resolution, and thus the computational results strongly depend on the numerical method, especially in simulation with poor resolution. In this paper, a new dynamic rocedure with the vector level identity, which takes the effect of grid resolution into consideration, was proposed. In addition, the dynamic Smagorinsky eddy viscosity model with the vector level identity was tested. All computational tests were done in the turbulent channel flow, and the Reynolds number based on the channel half width and wall friction velocity is 395. To remove the ambiguity regarding the accuracy of the finite difference scheme, the higher (up to 12th) order accurate fully conservative finite difference schemes in a staggered grid system proposed by Morinishi was used. The mean velocity profile computed using the new SGS model does not depend on the grid resolution confirmed.

View full abstract

Comparatively high Reynolds number channel flows up to Re_τ=640 are simulated by DNS. Noticeable property of the high Reynolds number flow is large-scale streaks which are larger not only in scale but also in their spanwise separation than those well-known in buffer layer. It is revealed that low speed streaks are the region of high turbulent activity and accordingly of high turbulent shear stress with densely populated elementary vortices. These properties are little influenced by wall condition, such as drag controlled or not. LSE (Linear Stochastic Estimation) technique is applied to extract elementary vortex pairs. Time evolution of these vertices submerged in a laminar flow but having velocity distribution of turbulent one reveals the process of generation of large-scale streaks and accompanying densely populated elementary vortices. The formation of large-scale streaks by the model simulation is also little influenced by wall condition and this modeled process is expected to mimic that taking place in the near-wall layer of practical channel flow. The existence of low speed streaks itself is the key for the formation of the structure, thus suggesting the self-sustenance of turbulence by large-scale streaks and vortices concentrated area.

View full abstract

If we take a no-slip boundary condition at the wall in LES, many grids are needed near the wall. But some wall models have been applied to decrease the calculation time for LES. In LES using the wall models, the wall shear stresses are calculated from them. This paper evaluates two types of the wall models, WM1 based on the standard logarithmic law and WM2 which solves the boundary layer equation (BLE) near the wall. WM1 holds well in the channel flow or the zero pressure gradient boundary layer (ZPGBL), however it is impossible to apply WM1 to the adverse pressure gradient boundary layer (APGBL);WM2 can involve the pressure effect because there exists the pressure derivative in BLE. In channel flow, both of WM1 and WM2 give a better prediction for the streamwise mean velocity than the LES with the no-slip boundary condition, though they need half grids as that of the LES with the no-slip boundary condition. In APGBL, WM2 predicts more successfully the wall shear stresses than WM1.

View full abstract

Large eddy simulation (LES) of compressible boundary layer including transition to turbulence was successfully done at Mach number of 0.3 and Reynolds number of 10⁵. Although there are many studies focused on incompressible turbulence, there has recently emerged the interest in LES for compressible turbulence. A compressible Smagorinsky-type eddy viscosity model was proposed and employed in this simulation. It was shown that the present simulation recreates well not only the time-averaged quantities but velocity fluctuations such as root-mean square of turbulence intensities. Also, the present LES gives 3D detailed structure of streak and bursting phenomena, which is difficult to simulate by using Reynolds averaged turbulence models.

View full abstract

Large eddy simulation (LES) of compression ramp flow including shock wave was successfully performed at a free stream Mach number of 2.9 and a Reynolds number of 10⁶ based on the boundary layer thickness. A compressible Smagorinsky-type eddy viscosity model was employed. It was shown that the present LES predicts not only the time-averaged quantities but fluctuations such as pressure deviation at compression corner. Also, the present simulation predicts three dimensional structure of shock/turbulent boundary interaction, which is essentially difficult to simulate by using Reynolds averaged Navier-Stokes equations (RANS).

View full abstract

A numerical simulation model was developed to predict the instantaneous concentration fluctuation of a tracer gas plume and applied to stack gas diffusion on a flat plate. The flow field, including an instantaneous velocity component, was predicted using the LES (large eddy simulation) method. Then, concentration was calculated using the turbulence closure method, in which the LES is expanded for concentration, and the puff method, in which small pieces of the tracer gas are divided and combined according to the calculation mesh sizes. In order to avoid numerical viscous effects, the puff method was applied in the regions near to the tracer gas release point. Numerical calculation results for concentration fluctuation were compared with those obtained in a wind tunnel experiment, and good agreement was obtained for high concentrations of probability distribution functions.

View full abstract

The paper presents a dynamic SGS model of Two Way Coupling for Large Eddy Simulation of particle-laden turbulent flow, which based on Yuu's model (1997). The advantage of this new model is that the coefficient of proposed SGS model can be decided by Germano's (1991) dynamic procedure. Then, the numerical simulation was performed with Van Driest wall function model and the present dynamic SGS model, for downward particle laden turbulent flows at Re=644 in a vertical channel that was experimentally investigated by Kulick (1994). The present dynamic SGS model has been verified through comparison of statistical properties of particle-laden turbulent flow. In addition, the effect of SGS component coupling on the turbulence modification has been investigated.

View full abstract

A large eddy simulation (LES) based on a finite volume method was applied to a liquid mixing layer flow downstream of a turbulence-generating grid with chemical reactions. The large eddy probability density function (LEPDF) model and joint LEPDF model were used as a SGS model for a rapid reaction and for a moderately fast reaction, respectively. The results of the LES were compared with the measurements to examine the proposed subgrid-scale (SGS) models. Furthermore, the correlation coefficient in the scale similarity model between the variance of SGS concentration and that of test-filtered concentration was determined using the data of the direct numerical simulation (DNS) of liquid isotropic turbulence. The results show that the correlation coefficient for a liquid flow 5 is times bigger than that for a gas flow. When the appropriate correlation coefficent is given, the LES results are in good agreement with the measurements.

View full abstract

A program for fluid and reaction analysis of supercritical fluid (SCF) is developed for the purpose of designing chemical reactors using SCF. In this program, the flow solution scheme with the CIP method is incorporated, which can solve the flow stable in which the density extremely fluctuates by hydrothermal reaction. The hydrothermal decomposition and oxidation of coffee residue in a SCF reactor are calculated as an example. The distributions of thermodynamic quantities such as temperature, density, enthalpy in the cylindrical reactor vessel are investigated in this study.

View full abstract

The situation called "fine granularity"is considered, which means that the number of control volumes or grid points allocated to one processor is small. Parallel efficiency usually deterirates in fine granularity, therefore it is crucial to develop a parallel Navier-Stokes solver of high efficiency. The authors have focused on the coupled method and applied it to steady flows on the Cartesian to show that its parallel efficiency was high. In this study, considering engineering applications, the coupled method is applied to unsteady flows and expanded to the boundary fitted coordinate. 2-D and 3-D unsteady flows past a body are calculated and high parallel efficiency of the coupled method is demonstrated.

View full abstract

Three-dimensional fluid flows are simulated numerically by the 3 D 15 V incompressible lattice Boltzmann model. In a rotating coordinate system, the effect of Coriolis force is incorporated into the local equilibrium distribution function at the collision stage. The force due to the pressure gradient is taken into consideration at the translating stage of the particles. The velocity fields are obtained by averaging these calculated two stage fields. The Taylor-Proudman column, the Ekman spiral and the spin-down are simulated. The existence of the Taylor-Proudman column is confirmed by the calculation. Results of the Ekman spiral and the spin-down agree well with theoretical ones obtained by an analysis with the boundary layer approximation.

View full abstract

Numerical study is performed on the vortex breakdown within a cylindrical container for the cases of spin-up from and spin-down to rest of an endwall disk. This flow problem is characterized by two non-dimensional physical parameters, i. e., the Reynolds number Re≡ΩR²/ν and the container aspect ratio H/R, here Ω, R, H and ν denote the angular frequency of the rotating endwall disk, the radius and height of the container, and the kinematic viscosity of the working fluid, respectively. In the present numerical investigation, the axisymmetry of the flow is assumed and the governing equations are solved by a finite difference method for the cases of H/R=2.00, 2.25 and 2.50, and Re=1.97×10³∼2.46×10³. Numerical results are discussed by comparing with flow visualizations. The appearance of oscillating vortex breakdown(s) and the damping to a steady state in the spin-up process and also the disappearance of vortex breakdown(s) and the appearance of a secondary recirculating zone in the spin-down process are demonstrated. Necessary condition for the appearance of the vortex breakdown suggested by Brown and Lopez is confirmed to extend for the spin-up process.

View full abstract

This paper proposes a two-dimensional numerical method for gas-solid two-phase free turbulent flow, which can take account of the interaction between the particle and the gas-phase. The computation of the gas flow by a vortex method and the Lagrangian calculation of the particle motion are simultaneously performed. The change in the vorticity of the vortex element by the viscous effect is simulated through a core spreading method. The change due to the particle is evaluated from the force exerted by the particle acting on the gas-phase using an area weighting method. The present method is also applied to the analysis of a gas-solid two-phase mixing layer. The numerical results, such as the number density, velocity and turbulent intensity of the particle, are favorably compared with the experimental date, demonstrating the applicability of the method.

View full abstract

It is well known that when a pressure wave generated by a high-speed train entering a tunnel is propagated through the tunnel and arrives at the distant tunnel exit, a micro-pressure wave is radiated from the exit portal. In addition, several pressure waves are radiated from the entrance portal when a train enters the tunnel and from the exit portal when the train leaves it. Here, the former is called a tunnel entry wave, and the latter a tunnel exit wave. Although the tunnel entry/exit wave is usually weaker than the micro-pressure wave, they could cause an environmental problem for future high-speed railways. In this study, numerical analysis is performed to clarify the phenomenon of the tunnel entry wave and predict its waveform. The numerical method is a combination of the unsteady compressible inviscid flow solver and the acoustic analogy. The numerical results show that a negative pressure pulse is radiated when the train nose enters the tunnel and a positive pressure pulse when the train tail enters. Also it is shown that the tunnel entry wave attenuates as the distance between the tunnel entrance and the observer point becomes larger but has a strong directivity in the direction of train running. The results of the numerical analysis in the cases where the entrance is unflanged and flanged are compared with those of model experiments and satisfactory agreement has been obtained.

View full abstract

The behavior of fine solid particles in a virtual impacter type classifier having a simple construction and the classification performance are investigated by both the numerical and experimental analyses. The applicability of the numerical analysis which based on the κ-ε two-equation model and the two-way method is testified by the experiment on the partial classification efficiency. It is shown that the result of numerical analysis and empirical data agree well with each other. The effect of the flow and structure parameter on the partial classification efficiency are analyzed by numerical methods for the classifier's control and optimization further more.

View full abstract

It is well known that the boundary layer separation, which is induced by the adverse pressure gradient of pseudo-shock, affects the dynamical feature of the pseudo-shock such as the shape of the first shock, the pressure recovery rate and the self-induced oscillation. So, it is considered that the suppression of the separation is effective on the shock wave/boundary layer interaction. In this research, the boundary layer control with the slot injection is studied experimentally and numerically. The expriments were carried out in a blow-down supersonic wind tunnel at a free stream Mach number of 2.0. Numerical simulations were made with the 2-D compressible Navier-Stokes code. In sloving the equations, 2nd-order accurate Harten-Yee's upwind TVD scheme and κ-ε turbulence model modifed for compressibility effects were used. The result shows that the separtion is appreciably suppressed and the first shock changes in its shape, so that the wall pressure fluctuation is greatly suppressed. In addition, the small slot high results in the lower total pressure loss.

View full abstract

The car-following model of traffic is extended to take into account the car interaction before the next car ahead (the next-nearest-neighbor interaction). The traffic behavior of the extended car-following model is investigated numerically and analytically. It is shown that next-nearest-neighbor interaction stabilizes the traffic flow. The jamming transition between the freely moving and jammed phases occurs at a higher density than the threshold of the original car-following model. The traffic current is enhanced without jam by the stabilization effect. The jamming transition is analyzed with the use of the linear stability and nonlinear perturbation methods. The traffic jim is described by the kink solution of the modified Korteweg-de Vries (MKdV) equation. The theoretical coexisting curve is in good agreement with the simulation result.

View full abstract

The interfacial phenomena of magnetic fluids subjected to a normal magnetic field are studied experimetally. To begin with, the effect of the shape and the dimension of the transparent containers on the interfacial phenomena is examined for two kinds of magnetic fluids, ferricolloid W-40 and ferricolloid HC-50. The cross-sectional shape of the container is made to be circular, square and hexagon. The dimension of the container is set to 55 mm, 70 mm and 85 mm. The critical magnetic induction values B_c calculated on the basis of present expriment are compared with those obtained from the theoretical analysis by Cowley & Rosensweig, where at these values of B_c the interfacial phenomena begin to be built up. Finally, on the hysteresis in generating the transition between 3 kinds of deformation modes on the interface, the present observation results are compared with those obtained from the analysis by A. Gailitis.

View full abstract

Recently, Considerable interest has developed in surfactant additives for use in district heating and cooling systems to lower pumping energy requirement. In this study, drag reduction for an enclosed rotating disk in surfactant solutions was clarified experimentally. Experiments were carried out to measure the torque of acting on one side of the disk. Test surfactant solutions were Ethoquad O/12 with sodium salicylate at concentration of 50, 100 and 200 ppm. The temperature of test solutions were 18, 25 and 31°C. It was shown that the drag reduction of the surfactant solutions was not only depended on the concentration but also on the temperature. From flow visualization technique, it has been clarified that the amplitude of circular vortices around the rotating disk were reduced by the surfactant solution and the flow direction near the disk was turned to outward. The maximum drag reduction ratio was about 30% in 200 ppm Ethoquad O/12 solution at about Re=5×10⁵.

View full abstract

Plants have many tee pipes where a hot fluid and cold fluid flow into each other and mix. The mixing of hot and cold water can cause a fluid temperature fluctuation which leads to the temperature flutctuation of the inner pipe wall, so characteristics of this flucuation in the mixing tee pipe were investigated through experiments. This paper shows the three mixing flow patterns in the tee pipe, i. e. stratified flow, turned jet and wall jet, the location of the maximum fluid temperature fluctuation and fluid temperature fluctuation characteristics downstream from the tee pipe. The experiments showed that characteristics of the fluid temperature fluctuation were largely related to the flow pattern in the tee pipe. The flow pattern of the turned flow in the tee pipe could suppress the fluid temperature fluctuation.

View full abstract

Cavitation occurring at a sharp projection in a hydraulic oil flow was observed, using a microscope, video-cameras, laser beam and a photo-multiplier. On a sewing needle employed as a projection, a tiny cavity as small as several 10 μm in length suddenly emerged at its tip and stayed afterwards. In vigorous cavitation the needle emitted light and simultaneously produced synchronized electric charge. At the same time the electrode placed downstream of the needle detected the electric charge of an opposite sign. The observations by the microscope made it possible to determine the exact start of cavitation, which turned out to depend on the oil temperature as well as the downstream pressure. Moreover the streamline stretching from the tip of the needle became visible as the flow rate increased. There was evidence that the highly viscous shear flow generated heat at the separation point, which probably explains why the streamline was visualized.

View full abstract

For turbomachineries operating at supercritical shaft speed, it is important to understand the characteristics of unsteady fluid forces on the impeller which occur due to shaft vibration. The present paper treats the forces on an open-type centrifugal impeller in whirling motion using unsteady potential flow theory. The whirling forces obtained agree fairly well with experimental results and show a destabilizing region at small positive whirl. It was found that the destabilizing force is due to the forces on the hub caused by the change in flow thickness, with minor effect of tip leakage on the destabilization.

View full abstract

This paper proposes a unique counter-rotating type of the pumping system which is composed of the two-stage impellers and the peculiar motor with the double rotors, and discusses remarkable advantages of the system. The front and the rear impellers are driven by the inner and the outer rotors, respectively, keeping the relative-rotational speed constant and keeping the counter-rotational torques same. Such operating conditions make not only the unstable performance in the low discharge but also the cavitation in the high discharge suppress smartly, in the optimum cooperation with the impellers and the rotors. The model test proved that there is not the rising portion of the head characteristics and the front impeller speed decreases with increase of the discharge as if the impeller takes the place of the inducer.

View full abstract

A simple method for predicting the mass flux of a turbomolecular pump with spiral grooves is proposed. The flow approaching the rotor is modeled. The flow is divided into two;one is the flow near the rotor and the other is the flow near the casing. The former is three dimensional and in the transition regime. It is analyzed using the Direct Simulation Monte Carlo method. The latter is treated as an averaged one-dimensional flow, which compensates for a shortage of the mass flux resulting from the former flow. The total mass flux can be obtained as a sum of the mass fluxes due to the two flows. The computational results agree very well with the measured data.

View full abstract

Natural convective flows over upward-facing, inclined plates were investigated experimentally. The attentions were focused on the opposing flows that will appear over the plates inclined slightly from horizontal. The flow fields over the plates and the surface temperatures of the heated plates were visualized with dye and liquid crystal thermometry. The result showed that both the descending and ascending flows appear over the plates when the inclination angle is less than 15°. These flows collide with each other at certain distance from the leading edge and, then, ascend away from the plate as a plume. It was found that the above distance is determined solely by the inclination angle and is independent of sizes and heat fluxes of the plate. The local heat transfer coefficients of the plates were also measured. The result showed that the heat transfer is enhanced slightly by the occurrence of the descending flows.

View full abstract

Natural covection heat transfer and flow structure in an anisotropic porous medium of square cavity saturated with Boussinesq fluid has been studied experimentally using a Hele-Shaw cell. The permeability ratio defined by K=K_y/K_x was put to three different values;0.4, 1 and 2.5. The convection patterns at three different permeability ratios are visualized for several different Rayleigh numbers by the pH indicator method. When K is 0.25, the visualized flow is mainly in the vertical direction. On the contrary for K=4 the convecting flow is in the horizontal direction. The average heat transfer coefficients are also measured, and the corresponding Nusselt number are plotted as a function of K. It is found that the corresponding Nusselt numbers are scaled with (KRa)^1/2. The experimental results of flow pattern and heat transfer are accord with those obtained by our previous theory.

View full abstract

Natural-convective heat-transfer in vertical concentric pipe annuli is investigated both numerically and experimentally for a fluid of Prandtl number of 0.7. Calculations for three cases different in the heating condition for pipes (heated inner pipe, heated outer pipe, heated both pipes) are made of laminar flows for varying the diameter ratio of inner to outer pipe d_i/d_o from 0.2 to 0.8. For each case, the length of thermal entrance region x/b is well correlated with the modified Grashof number Gr^* and the ratio of length to clearance of annulus L/b at the Grashor numbers Gr^*=10²∼5×10⁵. Local Nusself numbers at thermally fully-developed region show some constant values dependent on the diameter ratio d_i/d_o, regardless of Gr^* and L/b. Average Nusselt numbers for the thermal entrance region are also independent of Gr^* and L/b.

View full abstract

The present paper has dealt with a new defrost measure by using sublimation phenomenon which occurs below the triple point of water (273.16 K, 610.5 Pa) The present experimental study has focused to examine the mass transfer of annular frost layers developed on two cooling horizonal pipes set in a vertical direction which was exposed to an impinging jet air flow. The morphology of the frost layer during sublimating was observed by using a CCD camera. The nondimensional correlation equations of mass transfer were derived as a function of various parameters.

View full abstract

An energy storage system that stores energy in the form of liquid air is studied. In this system, the cool storage unit is the most important unit. From the viewpoint of safety and economy, it is most promising to store cold as sensitive heat of solids such as pebbles or concrete. To grasp temperature variation of the solid cool storage unit, a simulator was developed that calculates unsteady heat transfer between supercritical gas flow and solid material. Comparison of the calculation result with an experimental result showed that the temperature variation of the metal cool storage medium was accurate within 11%. For the concrete cool storage unit, the calculation result showed that with less pitch of the tube, a smaller quantity of medium was required. The minimum quantity was 3 times that of concrete, simply estimated from heat capacity of concrete and air. The volume required for concrete cool storage was less than 1/100 of that of reservoirs for a pumped-hydro power station having a drop of 500m.

View full abstract

The mechanism of suppression of vapor explosions by adding inversely soluble polymes in water are studied. The vapor exposion experiments and the cooling experiment of silver test piece are conducted. Polymetric solution, Poly (ethylene oxide : PEO), of concentrations from 0 to 500 wppm, whose viscosity relative to water ranges from 1.00 to 2.00, is used. No vapor explosion is observed in the polymer solution at a higher concentration than 200 wppm. Quench experiments of a silver test piece submerged in the polymer solution or water are performed in order to examine the stability of film boiling. The cause of suppression of the vapor film collapse is concluded be to due to the precipitation of PEO as a gel around the vapor film.

View full abstract

Using inverse radiative property value analysis, profiles of gray equivalent absorption coefficient are estimated which give results similar to the ones of nongray analysis when the coefficients are used in the corresponding gary analysis. For the inverse analysis, the results of nongray analysis are used as the input. A new approximate method is proposed for the estimation of the profiles of equivalent absorption coefficient without knowing tha results of nongray analysis previously. The profiles of the equivalent absorption coefficient obtained by the new method is compared with the ones estimated by the inverse analysis for six practical cases. And the new method is shown to be useful for estimating the profile of absorption coefficient for gray analysis in most cases.

View full abstract

A detonation-driven shock tube firstly designed by H.-R. Yu, is considered as a useful facilities capable of producing high-enthalpy flow. In this apparatus, a strong shock wave is generated by detonating oxygen-hydrogen (oxyhydrogen) mixture and has characteristics that temperature as well as pressure of driver gas is extremely high compared with conventional shock tubes. However, a structure of detonation wave is not uniform e.g., detonation wave has three-dimensional cellular structures and multiple transverse waves. Furthermore, the detonation wave is followed by a Taylor expansion fan and performance of detonation-driven shock tube is not well understood. In this preliminary study, a detonation-driven shock tube is constructed and its performance is experimentally investigated by measuring pressure histories and a profile of ionization current behind detonation wave. As a result, (i) the pressure histories of detonation wave is clarified and it shows reasonable agreement with a result obtained by KASIMIR shock tube simulation code. (ii) A propagation velocity of detonation wave is coincided well with theoretical predictions assuming Chapman-Jouguet detonation wave. (iii) An equivalence ratio of oxyhydrogen mixture to produce a highest Mach number of the shock wave is evaluated as 〓≃1.7.

View full abstract

Gas dissolution process with chemical reaction of a spherical rising gas bubble was investigated experimentally. We developed an experimental system that use a CCD camera coupled with a microscope to track the rising bubble. By measuring the bubble size and rising speed from the bubble motion data captured by a personal computer, we could precisely estimate the drag coefficients and the Sherwood number for the dissolution of gas bubbles when a carbon dioxide bubble dissolves in sodium hydroxide with chemical reactions. And we also estimated the dissolution rate of the bubble in alkaline solution using the numerically estimated dissolution rate in water when we assumed that the chemical equilibrium attained at the gas-liquid interface. We compared the numerical values with the experimental ones and the results show that the chemical equilibrium does not attain at the surface when the concentration of sodium hydroxide is higher than 0.1 mol/1.

View full abstract

Gas dissolution process with chemical reaction of a spherical rising gas bubble was investigated numerically. We numerically estimated drag coefficients and Sherwood number of the "Stagnant Cap Model" by directly solving the Navier-Stokes equation and the convection-diffusion equation when a carbon dioxide bubble dissolves in strong alkaeine solution with chemical reactions. In this analysis, we assume that the chemical equilibrium does not attain near the bubble is and include the chemical reaction term in the convection-diffusion equation. We compared the results with the measured rising speed and dissolution rate. The results reveal that the experimental and numerical results are in good agreement and the dissolution rate with the chemical reactions can be estimated even when the chemical equilibrium does not attain near the bubble. We also observed that the behavior of the bubble returns to that of a fluid sphere after it is closed to that of a solid particle when the concentration of sodium hydroxide is as high as 1 mol/1.

View full abstract

Rotational temperature of NO molecule in methane/air premixed flame was estimated by a spectral matching method. A tunable narrow band ArF excimer laser was used to excite the D²Σ⁺(ν′=0)←X²Π(ν"=1) band system of NO. Laser beam irradiated a flame, and the laser induced fluorescence was resolved into a spectrum by using a spectroscope. On this spectrum, ε and δ bands of upper vibrational level of ν′=0 were analyzed. In order to use a spectral matching method, profiles of ε and δ band spectra were calculated theoretically in detail with reliable molecular constants and exact formula, and they were modulated by an experimental slit function. Since the profile of band spectrum was determined as a function of a rotational temperature, a rotational temperature could be estimated from the temperature where the profile of every band spectrum obtained theoretically is fitted to that of experimentally obtained. Applying a spectral matching method on the ε(0, 3), ε(0, 4) and δ(0, 2) band of NO, it was obtained that the rotational temperature is about 1000K. Obtained rotational temperature is almost agreed with a thermocouple temperature.

View full abstract

In order to clarify the turbulent burning velocity of the multi-component fuel mixtures, the lean and rich two-component fuel mixtures, where methane, propane and hydrogen were used as fuels, were prepared keeping the laminar burning velocity nearly the same value. Clear difference in the measured turbulent burning velocity at the same turbulence intensity can be seen among the two-component fuel mixtures with different addition rate of fuel, even under nearly the same laminar burning velocity. The burning velocities of lean mixtures change almost monotonously as addition rate changes, those of rich mixtures, however, do not show such a monotony. These phenomena can be explained qualitatively from the local burning velocities, estimated by taking account of the preferential diffusion effect for each fuel component. In addition, a model of turbulent burning velocity proposed for the one-component fuel mixtures can possibly be applied to the two-component fuel mixtures by using the estimated local burning velocity of each fuel.

View full abstract

NO_χ emission was remarkably reduced by PREDIC (PREmixed lean DIesel Combustion) system in which fuel was injected at very early stage of compression stroke and the major part of the fuel is considered to be burned with self-ignition of premixed charge around TDC. However PREDIC system had some problems, a restriction of a high load operation was one of these problems. In order to investigate the combustion characteristics of PREDIC at the richer operation limit, a test engine was operated with gaseous fuel-air mixture where less heterogenoeus mixture can be formed than that of conventional diesel engines. A steep pressure rise or the abrupt increase in NO_χ emission determined the richer operation limit. This was at 2 to 2.4 of excess air ratio. Supercharging operation enabled the high load operation more than 2.4 of excess air ratio.

View full abstract

Thermal conductivity and thermal diffusivity of combustion chamber deposits in a diesel engine have been measured as fundamental data for heat transfer analyses. The deposits were built-up on the surface thermocouples under different engine loads, speeds and locations. After removing the sooted thermocouples from the engine, the thermal properties were determined by the combination of several techniques which were established in a previous study. The results show that thermal diffusivity is 2.0∼3.4×10^-6m²/s without clear dependencies on engine loads and locations. It is also shown that thermal conductivity is influenced by the engine loads resulting in 1.50∼1.84W/(m·K) at 80% load and 1.14∼1.41W/(m·K) at 60% load. This fact contrasts with the previous results for s.i. engine deposits whose thermal conductivity does not depend on engine loads. It is also found that that the thermal conductivity of the diesel engine deposits is larger in most cases than that of the s.i. engine deposits which is 1.06∼1.21W/(m·K).

View full abstract