In this paper, Free Mesh Method, which is a kind of meshless method based on the usual finite element method, is applied to unsteady incompressible viscous flow analysis using a parallel computer. The method is well compatible with the parallel environment because the global matrices of the finite element method, mass, advection, diffusion and gradient matrices, are independently computed node-by-node, and furthermore because the node-element connectivity information is unnecessary. To solve the governing equations of fluid dynamics, the Navier-Stokes and the the continuity equations, the velocity correction method with interpolation functions for velocity and pressure having equal order was used for time integration. A message passing library was used for the parallelization and the parallel efficiency over 83% was obtained for a large scale problem with 480 000 degrees of freedom using 16 processors.

View full abstract

The conventional shape function for the finite-element method (FEM) is a linear one, which is inadequate for analyzing numerically complex flows for high Reynolds numbers. In the present study, we suppose CIP-FEM scheme using 3rd order shape function which requires continuity of the value of the function and its first space derivative in whole space and is formulated a finite element method for the cubic interpolated pseudo-particle (CIP) method which was developed in the field of finite difference method. The CIP method is used to analyze Advection equations with high accuracy. We verify the effectiveness of this new scheme by 1-D and 2D-Advection Diffusion equations which have the exact solution. The numerical results of these problems are good agreement with each of exact one we could confirm the effectiveness of CIP-FEM scheme.

View full abstract

Three-dimensional swept shock wave/turbulent boundary layer interaction and its control by means of bleed are important subjects in supersonic air intake development. In the present study, the modeled flowfield is computed using an explicit numerical algorithm for the compressible Navier-Stokes equations. The results are obtained at free-stream Mach numbers of 3.25 and 3.11 employing zero and two equation turbulence models with and without boundary layer bleed around the interaction region. The results with two equation model show less agreement with the experimental data than those with zero equation model because of the overprediction for eddy viscosity. We confirm that the latter code is available for a tool to design a supersonic intake with bleed. Also, this study provides important information about the flowfield and reproducibility of bleed effect. For more accurate prediction, however, new modelling for bleed boundary conditions should be introduced.

View full abstract

Unsteady three-dimensional transonic viscous flows through a gas turbine cascade channel are numerically investigated using the higher-resolution shock-vortex capturing method proposed by the authors. The higher-resolution both in space discretization and time integration enables us to obtain real flow characteristics in the 3-D gas turbine cascade. As numerical results, calculated unsteady 3-D vortical structures in the wake and at the corner are shown, and obtained unsteady flow patterns near the blade under on-and off-design conditions are compared with eath other.

View full abstract

Fragmentation of liquid metal takes place as basic processes of vapor explosions. This leads to rapid evaporation on the stretched interface. To date, a number of models explaining the fragmentation mechanisms have been proposed. However, few evidence has been obtained from the experiment because the phenomena are rapid. In Moving Particle Semi-implicit (MPS) method, grids are not necessary so that fluid fragmentation as well as multi-fluid thermal hydraulics can be analyzed. A numerical model of evaporation is developed for the present study. Impingement of water jets on a liquid metal pool is analyzed using the MPS method to investigate two typical models explaining the fragmentation mechanisms : Kim-Corradini and Ciccarelli-Frost models. Penetration of the water jet, which is assumed in Kim-Corradini model, is not observed in the calculation results. A filament of the liquid metal is observed between two water jets as assumed in Ciccarelli-Frost model. The filament appears when the jet density is smaller than the pool density, while the penetration appears when the jet density is hypotheticaly larger. The usual combinations of densities in vapor explosions are in the region of Ciccarelli-Frost model.

View full abstract

This paper describes extensive computer-based analytical studies on the details of unsteady flow behavior around oscillating airfoils in turbomachinery. To consider the time-dependent motions of airfoil, a complete Navier-Stokes solver incorporating a moving mesh was applied, and the drag and lift coefficients for the cases of stationary airfoils and airfoils subjected to forced oscillation were examined. In the present study, a comparison was made between the two-and three-dimensional analyses for NACA0012 airfoils, and the separation vortex structure was examined in detail. From the numerical results, it was found that the separation vortex consisted of large-scale rolls with axes in the span direction, and rib substructures with axes in the stream direction. The three-dimensional analysis could simulate these rolls and ribs, but the two-dimensional analysis was inadequate to realize this vortex structure. This is the main difference between the two-and three-dimensional analyses. In addition, the formation of ribs was found to be affected by the forced oscillation, and the transformation or rolls increased and the vortex structure became fine as the oscillation frequency increased.

View full abstract

For a thermoacoustic refrigerator or combustion oscillation, the secondary effect of oscillation is a ruling factor. A Rijke tube is made by the same effect and its structure is very simple. It has these characteristics : heat is released by heat conduction from a solid heat source, there is a stable flow, and it is an open system. The effects of excitation and dissipation for a Rijke tube are evaluated quantitatively. An equation of acoustic energy conservation which is a ruling equation for local area is integrated, then each term is solved. Analytic solution, transfer matrix method for vibration mode, and numerical calculation for the area around the heat source are used. There is a big density change in the area around the heat source, so the method used is adapted accordingly. A prediction of oscillation growth rate calculated by numerical simulation is compared with the experimental result. They are the same order.

View full abstract

This paper presents a definite prediction method of crosion in a conduit, consisting of an elbow, due to the impingement of solid particles entrained in flowing water. Owing to high speed movement of the particles, the erosion can be assumed to be caused by deformation and cutting wear as discussed thoroughly by Bitter. To obtain the precise impinging velocities, the fluid flow is evaluated by using a widely adopted κ-ε tarbulence model and a two-way coupling Lagrangian method to take the mutual interference between the two phases into account. The predicted erosion of the elbow due to the particles agrees well with those observed. The predicted wear rate are also compared with the results of one-way turbulent and potential flows, to demonstrate the effects of turbulence and particlefluid interaction.

View full abstract

Wave shapes and flow fields of falling liquid films down an inclined plate have been simulated with a finite difference method in which the algorithm is based on the HSMAC method and free boundary conditions are treated with newly proposed method. Effects of inclined angle of the plate and surface tension of the liquid are discussed. In the case of a constant fluid property, a constant flow rate and a constant disturbance frequency, the variation of inclined angle changes the wave shape remarkably because it affects governing dimensionless parameters such as Froude number, Weber number, and dimensionless disturbance frequency. Effects of the Froude number and the Waver number are shown separately. Wave peak heights decrease with decreases of Froude number and Weber number. For lower Froude number and Weber number, circulation flow in a large amplitude wave becomes smaller. The wave peak height and the wave celerity of the simulation agree well to previously proposed correlation equations.

View full abstract

We propose a theory for flushing phenomenon of fine particles based on gas-solid two phase model and by using SOLA-VOF method calculate flow of gas and particle when an external load is applying to powder surface. In order to examine the flow near the onset of flushing phenomenon, we omit the effective stress since numerical calculation.

View full abstract

According to mathematical theorems on initial-value problem, real-valued characteristic roots are necessary for a system of partial differential equations. However, a one-dimensional, one-pressure two-fluid model has complex-valued characteristic root, so that it is called mathematically ill-posed. Up to the present, eigenvalue analysis of a two-fluid model has been conducted only for a simplified two-fluid model without any source terms. Although these simplified analyses can yield a necessary condition for well-posedness, they cannot give a sufficient condition. A necessary and sufficient condition can be obtained if and only if the eigenvalues of a complete two-fluid model with source terms are evaluated. Furthermore, numerical stability of the two-fluid model is also affected by source terms. The examination of source terms is, hence, indispensable to clarify the mathematical and numerical nature of a two-fluid model. In this report, the method to examine the mathematical nature of a complete two-fluid model with source terms is presented. Based on this method, it is revealed that the degree of ill-posedness of the two-fluid model is greatly affected by gravity.

View full abstract

It is known that dilute carbon black suspensions cause the drag reduction for flow around a sphere at Re>1.5×10⁵. In this study, the experiment was carried out to measure the first normal differences of water/fine solid particle suspensions which the suspended particle is about 1.3μm in diameter by using the cone-and-plate rheogoniometer. The experimental data of carbon black suspensions show the viscoelastic behavior. By extrapolating the value of the first normal stress differences, it was pointed out that the drag reduction is occurred by the viscoelastic effect of the dilute suspensions.

View full abstract

The unsteady behavior of a deformable bubble in an accelerating fluid is analyzed numerically. The numerical method is based on a finite-volume solution of the equations on an orthogonal curvilinear coordinate system. The effect of the acceleration of the surrounding liquid is investigated by introducing the periodic body force in momentum equations. Numerical results reveal that theoretical model gives good predictions of bubble behavior as far as the deformation is small. However, the deformation gives much more complicated behavior. In this study, the deformation effect is classified into two types, depending on the ratio of the deformation time scale to that of translational motion and forcing. One is the type in which a shape oscillation is excited directly from the oscillatory body force field, and the other is the one in which a deformation is caused by a hydrodynamic interaction through the translational motion.

View full abstract

Characteristic properties of flow in a two-dimensional symmetric channel with a suddenly expanded and contracted part are investigated numerically and experimentally. It is predicted that there occur three distinct types of flow in the channel depending upon the Reynolds number, which are symmetric steady flow, asymmetric steady flow and asymmetric time-periodic flow. The three types of flow are confirmed to occur experimentally by flow visualization technique. Pressure distribution along the center line of the channel is evaluated numerically. Especially, the pressure difference across the expanded and contracted part of the channel is investigated in detail. It is concluded that the pressure difference is not largely affected by the type of the flow but differs depending on Reynolds number.

View full abstract

An experimental investigation was performed on turbulent channel flow with injection through a slit. Turbulent intensities of velocity fluctuation and Reynolds shear stress were measured in detail by a hot-wire anemometer. The statistical properties, i. e., length scales of turbulence, skewness/flatness factors and turbulent. diffusions, were investigated. Turbulent intensity and Reynolds shear stress increase by injection, however, the correlation coefficient of Reynolds shear stress depends slightly on injection and is constant across the channel. The turbulent diffusions ((u²v)^^^-, (-uv²)^^^-) are negative in the near-wall region. The result reveals that the sweep event of bursting is strengthened by injection.

View full abstract

Utilization of the electromagnetic force realizes the control of the shape and the flow of the molten metals without contact with the surrounding wall surfaces, resulting improved quality of the products. In the present research the new method to deflect the liquid metal jet by means of a series of the skewed non-uniform DC magnetic fields was proposed, and the experimental and the theoretical studies were conducted on the deflection of the liquid metal jets. The deflection increases with the increase in the amplitude of the magnetic flux density and with the decrease in the wave length of the magnetic field and in the effect of the gravity.

View full abstract

Nanbu [Phys. Rev. E, 55-4 (1997), 4642-4652] found the law of Coulomb collisions in plasmas : A succession of small-angle binary collisions is grouped into a unique binary collision with a large scattering angle ; the law of scattering is given by the exponential cosine function. Here is proposed a Coulomb collision algorithm for weighted sample particles based on the above paper. We considered two cases ; one is the case when the weight of sample particles is different from species to species, and the other is the case when the weight of sample particles is different from particle to particle. Sample calculations show that the weight algorithm works well.

View full abstract

As the structure of water ring vacuum pump is very complicated, it is very difficult to measure the variation of gas pressure between impellers accurately. This is why the measurement of the variation of gas pressure between impellers has not yet been published in prior studies. Measuring the pressure variation of a gas between impellers is very important in order to understand the characteristics of water ring vacuum pumps. We conducted an experiment to determine how gas is compressed between impellers in a water ring vacuum pump, and to determine the pressure variation of gas from the start of suction to the discharge of a gas. In the previous report, we demonstrated that the sealing water supply has a substantial effect on the characteristics of the water ring vacuum pump. We also analyzed how the sealing water supply affects gas pressure, volumetric efficiency and the quality of the vacuum.

View full abstract

This paper describes the relationships among rotation speed of runner, setting angle of blade, chord length, number of blade, chord length, number of blade, and the necessary swimming velocity of fish which pass through the runner theoretically. With fish experiment, the swimming velocity of fish is examined by many video films. Comparing with the data of theoretical prediction and experiment, in non-load operation, both are consistent. But in load operation, the fish velocity of experiment is less than that of theoretical prediction. The main reason is that most of fishes which passed through the runner are touched by blade. Fishes spend the time to swim away from the blade after touching the blade. We also investigated the hit position of fish body. In the condition of rotation speed below 200 rpm, except the head and the upper half of the body, the damage and mortality of fish will not happen. But the actual swimming velocity becomes slower. The higher is the rotation speed and the less is the blades number, the faster is the fish swimming velocity. The fish goes through the runner with the best speed which they decide and they choose the easiest course in the passage and regulate the swimming velocity themselves. The results supply the data for the study of biology co-exist water turbine.

View full abstract

The Darrieus-type cross-flow turbine, set in the duct casing, has been proposed for utilizing water generation at low head site in rivers and conduits. A weir existing anywhere in river might be replaced with this turbine system, flow passage of which consists of an intake, the runner section and a draft tube. Influences of runner casing geometries and draft tubes with various diverging angles on the turbine efficiency, that is a ratio of the generating power to supplied water head power, were investigated experimentally. In the present paper, measured velocity profiles in the passage are shown and preferable geometries of the casing and draft tube are discussed for effectively utilizing low head water power. And finally guiding principles for its design with efficiency of 57%, is demonstrated with useful data.

View full abstract

The design of the two-phase flow loops applied for thermal control systems of future spacecraft requires knowledge of two-phase flow and heat transfer phenomena under microgravity conditions. This paper deals with a fundamental approach for evaporators or cold plates in the two-phase flow loops. A straight evaporator tube was mathematically modeled as a simplified thermal configuration of the cold plate. Simultaneous ordinary differential equations for the two-phase annular flow under microgravity conditions were derived from the separated flow model, and numerical analyses were conducted for the purpose of parametric studies. As a result, properties' distributions of the working fluid along the flow direction, critical limits of the working conditions, and the thermal performance difference owing to the working fluids were clarified.

View full abstract

Rewetting of vertical-hot-thick slab was investigated experimentally and analytically. Quenching tests of thick copper and brass blocks were conducted for a falling film of saturated water in a range of the flow rate from 0.096 to 0.639 kg/ms at atmospheric pressure. The test sections were 50, 25 and 230mm in width, thickness and height, respectively. The rewetting velocities of the oxidized surfaces, which corresponded to quenching velocities, were faster than those of the non-oxidized surfaces irrespective of the flow passage spacing and the surface roughness. The rewetting velocities of the brass block were lower than those of the copper block. Relations between heat fluxes and wall superheats, which were obtained by solving a two-dimensional transient inverse-conduction equation, indicated that the heat transfer characteristics during the rewetting were quite similar to those of the conventional pool boiling. Simulation of the rewetting was also performed by using the two-dimensional transient heat conduction model with the boundary conditions corresponding to the pool boiling characteristics. The rewetting velocities measured were well reproduced by the simulation. The experimental and the simulation results suggested that precursory cooling during the film boiling, i. e. the stored energy just before the wetting, and the transition boiling heat flux just after the wetting had a large effect on the rewetting velocity ; the smaller the stored energy was and the larger was the transition boiling heat flux, the faster the revetting vecolity.

View full abstract

Analogy of convective heat transfer between developing laminar flows in curved pipes and orthogonally rotating pipes is described through similarity arguments and numerical computation. Governing parameters and a dimensionless axial distance are properly used for the respective flows. When the second parameter is large in each flow, it is shown that the temperature profiles and the Nusselt numbers of the two flows are approximately similar for the same values of the governing parameter, Prandtl number and the dimensionless axial distance.

View full abstract

The characteristic of boiling in a horizonatal channel with changing conditions of heated wall length and channel height have been studied experimentally. Behavior of bubbles on the heated wall and growth of bubble in the channel were observed by high speed camera at the experiment time. As a result, the behavior of growth of bubble which was classified into three types according to channel height had influence on time variation of degree of superheat, heat transfer and burnout heat flux in the channel. When the thickness of liquid on the bottom wall became thin, nucleate boiling with vapor dome was observed on the heated wall.

View full abstract

A new explicit algebraic gradient diffusion model for turbulent scalar (heat) flux in forced convection regimes is proposed. This model complies with the linearity and independence principles for passive scalar. For satisfying these, the model coefficients/functions do not include any turbulent scalar variable such as the temperature variance. The model thus does not require solving transport equations for turbulent quantities of the scalar field. The basic role of each model coefficient has been optimized by referring to the discussion with the LES data reported by the authors. The proposed model is applicable to a wide range of passive scalar fields by being coupled with a nonlinear κ-ε-A three equation turbulence model for dynamic fields. It has been confirmed that the model is able to capture the components of turbulent heat flux vectors in wall shear flows at least in the range of 0.025⪈Pr⪈95. Moreover, it has successfully distinguished the heat flux characteristics near shearfree boundaries.

View full abstract

Dry-out heat fluxes were investigated experimentally for a film flow falling down on the inner surface of vertical heated-tubes and for a low mass flux forced-upward flow in the tubes using R 113. This work followed the study on those for a two-phase natural circulation system. For the falling film boiling, flow state observation tests were also performed, where dry-patches appearing and disappearing repeatedly were observed near the exit end of the heated section at the dry-out heat flux conditions. Relation between the dry-out heat flux and the liquid film flow rate is analyzed. The dryout heat fluxes of the low mass flux upflow are experessed well by the correlation proposed in the previous work. The relation for the falling film boiling shows a similar trend to that for the upflow boiling, however, the dry-out heat fluxes of the falling film are much lower, approximately one third, than those of the upward flow.

View full abstract

Effect of heated length on the Critical Heat Flux (CHF) of Subcooled Flow Boiling with water was experimentally investigated by using direct current heated plate made of stainless steel with various lengths. The motion of bubbles at near burnout was ovserved by high speed video camera under atmospheric pressure condition. Consideration was made for critical wave length of Helmholts instability in several CHF mechanistic models. The slug length at CHF was almost coincided with the critical wave length. It is further confirmed that when the heated length is shorter than the critical wave length, the large enhancement of the CHF is shown.

View full abstract

Effect of heated length on the Critical Heat Flux (CHF) of subcooled flow boiling with water was experimentally investigated by using direct current heated tube made of stainless steel a part of whose wall thickness was axially cut for realizing nonuniform heat flux condition. The higher enhancement of the CHF was derived for shorter tube length. The effective heated length was determined for the tube under axially nonuniform heat flux condition. When the lower heat flux part below the Net Vapor Generation (NVG) heat flux exists at the middle of tube length, then the effective heated length becomes the tube length downstream the lower heat flux parts. However, when the lower heat flux part is above the NVG, then the effective heated length is full tube length.

View full abstract

Transient cooling of water around two cooled horizontal cylinders placed in a rectangular enclosure has been studied. The governing equations were solved by a finite difference method, and the flow structure and temperature distributions have been predicted. The purpose of the present study is to examine the effects of the spacing between the cylinders and the initial temperature of water on the cooling process of water. The initial water temperature was varied at 4, 6, 8 and 12°C, and timewise variations of the temperature field and the velocity field as well as the mean Nusselt numbers on the cylinder surfaces were compared. The effect of a change in the spacing between the cylinders on the cooling process appears earlier stage as the initial temperature of water is increased. Complicated flow patterns are observed for initial water temperatures larger than 4°C due to the density inversion of water, irrespective of the extent of the spacing.

View full abstract

Gas flow and heat transfer in a horizontal flow channel partially heated from below under 1G and reduced gravity have been studied experimentally and theoretically. Flow visualization and temperature measurements were conducted using short term microgravity facilities, i. e. a drop shaft and parabolic flights. Transient behavior of the gas flow patterns in accordance to the time dependent gravity changes were visualized by means of the suspended TiO₂ particles, illuminated by a laser sheet. Observed flow patterns showed fairly good agreement with the results of three dimensional unsteady numerical simulations. The criteria for the onset of transverse and longitudinal roll cells were mapped on Grashof number (Gr) vs. Reynolds number (Re) plane and well explained by the correlation lines previously reported by the authors based on steady two dimensional numerical simulations.

View full abstract

Experiments were performed to investigate the friction pressure drop of upward air-water two-phase flow in vertical twisted-tape-inserted tubes. Total pressure drop and average void fraction for hydrostatic pressure drop were measured at the same time, from which the friction pressure drop was calculated. The friction pressure drops of twisted-type-inserted tubes were 1.2∼5 times as large as that of empty one. However, the two-phase friction multiplier of the former was smaller than that of the latter and decreased with decrease of twist ratio (more tightly twisted). Such trends were predicted well by the well-known Chisholm's equation with modified value of the constant C based on the present experimental data. A correlation for the constant C was obtained as a function of twist ratio, with which all the present experimental data of friction pressure drop in the twisted-tape-inserted tubes were predicted well within ±30 percent error range.

View full abstract

Precise streaklines obtained by numerical calculation of the laminar, time-dependent continuity, Navier-Stokes and energy equations in an opposing flow of air show that the vortex scale increases, the vortex spiral is activated, and the wake frequency decreases, with an increase in the Richardson number Ri at the given Reynolds number Re. The increase of the vortex scale in an opposing wake was also shown by the present smoke-wire experiment. Thus the isothermal Karman vortex street develops due to negative buoyancy, and the developed Karman vortex street occurs in the wide range of Ri. These are resulted in the generation of wake vorticity due to negative buoyancy. The vortex street developed at the given Re and Ri never occurs in an isothermal wake, a wake with negative buoyancy at a low Re, and a wake developed at other Re and Ri.

View full abstract

In order to measure concentration in gas flow, a probe which consists of a sonic nozzle and a hot-wire was developed. The concentration probe of this type was named hot-wire concentration probe. We already proposed a method of concentration and velocity measurements by the hot-wire concentration probe and an ordinary hot-wire probe, and simultaneously measured instantaneous values in the turbulent jets. The present paper deals with another method by using the hot-wire concentration probe, a piezoresistive sensor, and an impact tube. Instantaneous velocity in gas mixture is given from density measured by means of the hot-wire concentration probe and total pressure measured by the piezoresistive sensor and the impact tube. In the pulsating jet where carbon dioxide gas pulsed by a loudspeaker issues into still air, the concentration and the velocity are measured by the above two different methods. It is confirmed that the measured values by both methods agree well. The experimental results indicate that the pulsation enhances mixing of two gases. Also, concentration and velocity changes do not occur at the same time.

View full abstract

The present paper deals with solidification characteristics of carbon dioxide around gas feed nozzle of the flue gas from thermal energy power plant. The present experiment was carried out under condition of dioxide content for 3∼50 volume percents. The results obtained indicated that the solidification fraction of carbon dioxide increased with a decrease in cooling temperature. The solidification fraction of carbon dioxide was expressed as the function of non-dimensional temperature and concentration of carbon dioxide.

View full abstract

A Monte Carlo method is presented to simulate the diffusion process with adsorption in porous media. The result of Monte Carlo simulation is compared with that of the difference equation method. It is shown that the Monte Carlo method is more efficient than the difference equation method. The effect of configuration of the adsorption particles upon the total concentration of molecules is investigated by the use of the Monte Carlo method. It is found that the configuration of the adsorption particles has an important effect on the diffusion and adsorption processes. Also, the effect of the adsorption rate upon the total concentration is studied.

View full abstract

The ram accelerator is a device to accelerate a projectile up to hypervelocity in a tube filled with combustible gaseous mixtures. One of the most troublesome problems in operating the facility is a difficulty to start. In order to improve the ram accelerator system in Hiroshima University, direct numerical simulations are carried using finite difference methods. A projectile and an igniter move at speed of 1 200 m/s from an evacuated chamber into a ram acceleration tube, where hydrogen-oxygen-nitrogen gas mixtures are filled as combustible gases. When combustible gas mixtures are filled in the acceleration tube, any solution without detonation or quenching is not obtained. To solve this problem, we introduce a ignition tube between the evacuated chamber and the acceleration tube. In the ram acceleration tube undetonable but combustible gas mixtures are filled, while in the ignition tube detonable gases are used. As a result, a solution is obtained with a choking mode combustion when the boundary layer is taken into account.

View full abstract

Direct numerical simulation of a H₂-Air turbulent premixed flame propagating in two-dimensional homogeneous isotropic turbulence was conducted to clarify local flame structure of turbulent premixed flames. Detailed kinetic mechanism including 21 elementary reactions and 8 reactive species was used to simulate H₂-Air combustion in turbulence. Local heat releases are well correlated with curvatures of the flame elements. Flames convex toward the burnt gas show large heat release, while those concave toward the burnt side show relatively low heat release. In the range of small curvatures, local heat releases increase linearly with the increase of curvature. Local heat release can be represented as a linear function of strain rate tangential to the flame front. Mean tangential strain rate at the flame fronts shows positive value, which results in the larger mean heat release than that of the laminar flame. The change of local heat release in turbulence can be explained by the modification of local flame structures due to difference of diffusion coefficients of O₂ and H₂.

View full abstract

The color of laminar diffusion flames burning propane, methane and ethylene was determined by chromaticity coordinates (x, y) defined by the CIE 1931 standard colorimetric system. The differences in flame color attributed to burning condition and fuel types were examined with a colorimeter. Spectroscopic measurement and numerical analysis using a simplified radiation model were also carried out to discuss the determining factors of the flame color. The relation between x and y measured on the central axis of the flames was expressed in the experimental equations. The (x, y) in the luminous region plotted on a chromaticity diagram changed along Planckian locus with the burning conditions. The contribution of the thermal radiation of soot particles and the chemiluminescence to the flame color was successfully evaluated by introducing the concept of additive mixture of color stimuli. The (x, y) profiles from the numerical analysis agreed well with the experimental results.

View full abstract

The effects of rotation on the stability and structure of tubular flames have been investigated for a lean hydrogen/air mixture, and mechanism of flame stability in a rotating flow has been discussed. Results show that with increasing intensity of rotation, (1) the flame diameter increases, while thickness of the luminous zone decreases, (2) the fuel concentration at the extinction decreanes, i. e., the stable region expands, (3) the radial temperature distribution becomes an M-shaped profiles, and (4) flame temperature T_f increases near extinction. These results are almost the same as those obtained for a methane/air mixture. The mechanism for these phenomena is supposed to be explained with the coupled effects of pressure diffusion and stretch effect. The present experimental results provide very useful information for discussing the flame stability and the transport phenomena in a vortex flow.

View full abstract

The velocity variation induced by the sinusoidal wall oscillation is measured along the stagnating streamline and its amplitude is estimated. The mechanisms of the flow velocity oscillation are discussed in terms of the amplitude of the flow velocity oscillation when the angular frequency of the wall oscillation ω, the average distance between the nozzle and the wall y^^-_ω, and the flow velocity at the nozzle exit U₀ are varied. The characteristics of the flow velocity oscillation are classified into three regions by a Strouhal number S=ωy^^-_ω/U₀. When S<2, the flow velocity variation follows the change of the wall location in a quasi-steady state. On the other hand, when S>6, it is dominated by the change of the wall velocity. In between, 2<S<6, the transition region is observed.

View full abstract

Occurrence behavior of microexplosion in droplet combustion of fuel mixtures is studied. Experiments were performed using unsupported droplets of n-alkane/n-hexadecane mixtures in normal gravity. It was found that the occurrence of microexplosion is stochastic and cannot be predicted by the classical criterion for microexplosion occurrence using the limit of superheat and the droplet temperature. An occurrence model for the microexplosion based on the homogeneous nucleation theory is presented and shows that the occurrence probability of the microexplosion is determined by the ratio of the liquid phase lifetime to the nucleation time during the quasi-steady vaporization period. The nucleation time is inversely proportional to the nucleation time is inversely proportional to the nucleation rate and superheated liquid volume. The relative value of the droplet temperature and the limit of superheat affects the occurrence probability through the nucleation rate. It can be well explained by the model that the occurrence probability has the maximum value for a certain initial concentration of the fuel mixture.

View full abstract

Experimental studies of the catalytic ignition of hydrogen-oxygen were performed. A spherical platinum of 1.5mm in diameter was used for catalyst and the atmospheric pressure was 0.1MPa. The parameters were equivalence ratio and dilution ratio. Dilution ratio is mole fraction of nitrogen. The surface temperature was measured with thermocouple and the time histories of surface temperature show an inflection point. The catalytic ignition temperature is defined as the temperature at the inflection point. The catalytic ignition temperature was increased with dilution ratio when equivalence ratio was constant. The effect of dilution ratio was explained well with adsorption rate of reactants. If equivalence ratio was less than 0.1, the catalytic ignition temperature was decreased with equivalence ratio increase, and was increased with equivalence ratio if equivalence ratio was more than 0.2. The desorption of adsorbed hydrogen should be considered to explain the effect of equivalence ratio.

View full abstract

In order to examine the effect of entrance and exit areas on the velocity distribution in the stacked wire gauzes which are the most familiar regenerator matrices of Stirling cycle machines, the pressure drop across it and the velocity distribution at its exit have been measured for different mesh numbers, flow rates, entrance areas, exit areas and stack thicknesses. The results show that the effective flow area ratio which is an index for the uniformity of the velocity distribution is independent of the mesh number and the Reynolds number but dependent on the entrance and exit areas and the stack thickness. In addition, a simple theoretical flow model has been led based on the isotropy of the friction factor which is confirmed by the measurement of axial and radial friction factors. The calculation using this model brings such an effective flow area ratio and a velocity distribution as agree well with experimental ones.

View full abstract

Measuring the torque of any rotational mechanical systems is very useful not only for measurements but also for controls. However, the cost of torque meter is the biggest practical problem. Therefore, computerized software torque meter was attempted making automobile wheel axle provided a gasoline engine an object. The theory applied here is not conventional Luenbergertype one but adaptive observer one by Luders-Narendra because of overcoming time-varying characteristic as well as non linear one of the objective system. Experimental results showed that statical precision was less than ±15 Nm, and dynamic response time was about 2 seconds. The latter, dynamic response is not enough as an automobile use. So, as a countermeasure, the concept of software torque meter with table-looking up method which modified by this adaptive observer, was also proposed.

View full abstract

Spatial characteristics of fine droplets formed at an initial injection stage has a great influence on mixture formation and ignition of diesel sprays. In this paper, a size and shape of each droplet near a nozzle tip has been analyzed by the back light illumination photography using a nano-spark light technique. The results show that fine droplets are formed immediately after 50μs from the commencement of the injection near the nozzle tip regardless of different nozzle hole sizes. Then, at the early stage of the injection period, the structure like a branch occurs at the periphery of a dense spray behind a spray tip and a number of fine droplets were produced in this region.

View full abstract

This study was performed to clarify the mechanism of mixture formation at peripheral area of diesel spray with PIV technique. Two dimensional cross-sectional photographs of diesel spray were taken with double pulse laser sheet. Local fuel spray particles were analyzed with an auto-correlation method and velocity vector and vorticity of the fuel spray particle were obtained. The vortex number increased and vorticity scale became smaller and its value grew higher with both smaller injection nozzle diameter and higher fuel injection velocity. With this injection condition, the mixing of fuel spray with ambient gas seems to be improved and the turbulence is expected to increase in the regions of higher vortex number, higher vorticity and smaller vorticity scale. Based on above results, the branch-like structure of diesel fuel spray was considered to be caused by vortices which formed in the shear layer between the spray and the ambient gas.

View full abstract

To evaluate the effectiveness of low heat rejection engine under heat loss condition, instantaneous heat fluxes flowing into ceramic piston surface and aluminum alloy (Loex) piston surface using thin film thermocouple were measured, and both were compared. As a result, in the working stroke, the instantaneous heat flux flowing into ceramic piston surface was larger than the instantaneous heat flux flowing into Loex piston surface. Accordingly, it became clear that reduction of heat loss was not effected when ceramics that thermal conductivity is small was used for combustion chamber wall.

View full abstract

In order to study the effect of fuel injection rate shaping on spray combustion, an electronically controllable fuel injection system was developed. In this system, the arbitrary injection rate shaping was obtained by controlling the movement of nozzle needle directly with piezoelectric actuator while the fuel pressure was maintained its constant injection pressure. By using this injection system, fuel was injected into the pressure chamber with three types of fuel injection rate shaping with different slope of injection rate rise. From the results of visualizations of non-evaporating spray, it was found that the slope of injection rate rise made great influences on the spatial and temporal fuel distributions and spray temporal evolution. Moreover, it was shown that, though there was no great difference in spray angle in low injection rate, spray angle by lift control injection was larger than that by pressure control injection in high injection rate.

View full abstract

Newly developed Stratified Fuel-Water Injection System was tested for the NO_x reduction on high-speed diesel engine. Compared with ordinary emulsified fuel oil combustion, this system enables to confirm the auto-ignition characteristics under the high water content condition, by means of primary injected pure fuel. Test results show that not only the broad reduction of exhaust NO_x, but also the improvement of B.S.F.C. and smoke density are achieved simultaneously. They also indicate that the exhaust NO_x in high-speed diesel engine could be less than 100 ppm (Excess O₂=13%).

View full abstract

An Ocean Thermal Energy Conversion (OTEC) system using a new cycle with absorption and extraction processes is proposed. In this paper, parametric performance analysis of OTEC system using the new cycle is carried out. The thermal efficiency of the OTEC system using the new cycle is compared with that of the Kalina cycle. Thermodynamic physical properties of ammonia/water mixtures are calculated using an improved version of the modified Benedict-Webb-Rubin (BWR) equation of state.

View full abstract