A hydrogen storage system using metal hydride (MH) has several problems to be solved before practical use. Among of them a long charging time required due to the poor heat transmission in MH bed during exoergic hydride forming reaction is essential for the hydrogen storage system of fuel cell electric vehicles. Four small tanks (effective hydrogen capacity 1.25 Nm³) using a La-Ni based AB 5 type hydrogen storage alloy were made by way of trial to attain the charging time within 10 minutes absorbing 80% of effective hydrogen capacity. The experiments were carried out to evaluate effects of thermal design of the tank, coolant condition on hydrogen absorption rate. Calculation of the process was done to improve performance of heat transfer in MH bed with additional thermal fins and rearrangement of coolant channels. The final version of the tank satisfied the required charging time even for higher coolant temperature of 21°C which reduces chiller load of a hydrogen station.

View full abstract

Two-dimensional direct numerical simulations have been performed in order to investigate the surface waves of falling films for three kinds of working fluids by using MARS (Multi-interfaces Advection and Reconstruction Solver) method. The surface waves can be controlled by an external artificial velocity oscillation. Although the amplitude of the artificial velocity oscillation is very small, the disturbances create small surface waves and eventually they develop to large surface waves consisting of solitary waves and capillary waves due to instability of falling films. The numerical results indicate that the surface-wave patterns of these two-dimensional waves can be divided into three groups with the dimensionless frequency, the ratio of the Weber number to the Kapitza number. It is found that these surface wave groups don't depend on the Reynolds number. According to the present results, the new correlations for the wave peak height with the Kapitza number and the dimensionless frequency and for the wavelength with dimensionless frequency are proposed. These correlations for three kinds of working fluids used in this study are shown in good agreement with the numerical and experimental results in a range of Re=30-400.

View full abstract

Three-dimensional turbulent flows past a square cylinder are simulated by the Finite Difference Lattice Boltzmann Method (FDLBM). We carry out two sort of approaches, i.e., first case is carried out by fourth order nurmerical viscosity without any turbulent model and second case is simulated using the Dynamic Smagorinsky Model (DSM). These results are compared with those by the experiments conducted by Lyn et al. Numerical results by DSM agree with experimental ones and it is confirmed that this method is useful for numerical simulations of turbulent flows.

View full abstract

The objective is to make clear the mechanisms of heat transfer under the bouyancy effect. A Direct Numerical Simulation (DNS) of natural convection in differentially heated vertical slot is conducted with Boussinesq approximation. The discretization of computation is carried out with a spectral method for three directions. Rayleigh number (Ra) which is defined with the slot width and the temperature difference between two walls, is 5.4×10⁵. Turbulence structures are investigated in more detail compared to previous researchs. As a results, it reveals that quasi-streamwise vortices are not formed in the near-wall region, the large-scale streamwise vortical structure is dominant in the core region of the slot. Then we extracted the large-scale structure by means of proper orthogonal decomposition (POD). Most energetic mode is characterized by the rotational motion, whose axis is in the streamwise direction and its scale is approximately the width of flow passage, and the reconstructed flow structures with low-dimensional modes of velocity highly correlates to those of thermal field. Therefore these findings recommend the control of the streamwise directed vortex to enhance effectively the heat exchange.

View full abstract

To perform massive and unsteady calculations of flow fields, the hybrid RANS/LES method is very useful, in which a RANS is employed near the wall, and an LES is carried out in the region away from the wall. However, to calculate adequately near-wall turbulent motions over complex terrain, a number of grids are needed near the wall. Thus, to reduce the calculation load when predicting real flow fields by computer, solving fewer transport equations is recommended. Therefore, in this study, we construct a low-Reynolds-number (LRN) one-equation model for eddy viscosity by introducing a time-scale based on the velocity gradient parameter. This time-scale has proven very effective for predicting local flows over complex terrain. The proposed model is evaluated in complex turbulent flows, such as backward-facing step flow, forward-facing step flow, flow past a two-dimensional bock and flow over a curved hill. The predicted results are compared with available experimental data. It is shown that the proposed one-equation model reproduces the results generally in good agreement with experimental data.

View full abstract

An experimental study is performed to investigate the position of falling liquid film apart from downward-facing surface for impinging jet. Besides, we investigate the circular hydraulic jump radius on upward-facing surface to compare with the radius of falling liquid film apart from downward-facing surface. The results show that both of the two radii have the same dependency with variation of tube Reynolds number. Water columns in dropping region are formed concyclic at regular intervals. It is substantiated that the interval is approximately the critical wavelength of Rayleigh-Taylor instability.

View full abstract

Influence of adding cationic surfactant (Ethoquard O12) with a consentration of 50 ppm on water flow around a cylinder in a range of Reynolds number Re_s between from 64 to 600 was investigated by using LDV and PIV. At Re_s =300, the flow with the surfactant formed a stagnant zone, which appeared periodically, in front of the cylinder expanding along the axial direction. The period of the appearance was about 120 [s]. The periodic flow structure was visualized by the measurements of velocity fluctuation. The development of the stagnant zone in front of the cylinder induced the expansion of a low speed flow area near the side of the cylinder.

View full abstract

Ultrasonic Velocity Profiler (UVP) was applied to an environmental flow attempting an efficient measurement of river flow. In a lab-scale open channel, a velocity distribution in a channel cross section was obtained and a flow rate was estimated. We confirmed that the flow rate in an open channel can be measured with an accuracy as high as 5%. It was also established to detect the bottom position from a single velocity distribution, which might give us information about the geometry and condition of the bottom surface of the channel. We practiced the method in a small river, and showed that it might give us flow rate of the river with higher accuracy than conventional methods.

View full abstract

In order to measure local blood velocity in artery and vein, a miniaturized fiber-optic LDV sensor has been newly developed. In the present fiber-optic sensor, laser beam emitted from the fiber tip can be focused at any position from about 0.1 to 0.5 mm distant from the fiber tip, and moreover a stray light or an unfavorable reflected light inside the light transmission system is remarkably decreased. Consequently the present sensor has a sufficiently high signal-to-noise ratio, and is capable of measuring the local velocity in almost all semi-opaque and opaque fluids including blood. The sensor head, i.e. the optical fiber is incorporated into either a injection needle or a catheter tube. In the former case, it is inserted inclinatorily from the wall into the vessel, and is made a tilted scan across the vessel to measure a velocity distribution. In the latter case, it is kept parallel to the flow in the vessel. In all the flows of whole human blood, whole caprine blood and a RC-MAP (red cell / Mannitol-Adenine-Phosphate) of bovine, of which hematocrit is 69%, the volocity distribution across the vessel can be very accurately measured. The insertion angle of the fiber to the flow has a significant effect on the measurement accuracy, and the appropriate angle is about 100 degree in the injection needle type. In the catheter type, the insertion direction adverse to the flow is better than that parallel to the flow because of an effect of wake behind the fiber tip. The influence of light source, i.e. a He-Ne gas laser and a He-Ne laser diode on the measurement accuracy is small.

View full abstract

Static pressure fluctuation is an important ingredient in turbulence, however the variation of the pressure fluctuations inside turbulence has not been studied so far. One reason for this is the difficulty inherent in measuring the fluctuating pressure and, consequently, a lack of published experimental results of this quantity. In this paper, pressure fluctuation is measured in turbulence using a condenser microphone and piezoresistive transducer. We summarize the matters to be attended to careful measurements and present a new method to remove the effect of resonance and standing waves. In order to check the accuracy of measured data, turbulence energy balance in a wake of cylinder is computed. Also, the statitical quantities, such as spectra and probability density function are compared with DNS.

View full abstract

This paper describes the fluid force acting on circular cylinder of finite length in the locking-in region. The experiment was carried out in an N. P. L. blow-down type wind-tunnel with a working section of 500 mm×500mm×2000mm at Reynolds number of 1.32×10⁴.The cylinder was forced to oscillate sinusoidally in the lift direction. The power spectrum of the fluctuating velocity in the wake behind a circular cylinder was measured to show the locking-in region in the present experiment. The time-mean pressure distribution and fluctuation pressure distribution on the circular cylinder were measured for the displacement in the vibration. Consequently it was found that the mean drag and fluctuating lift increase and become maximum in the locking-in region, while the base pressure in the rear surface (θ =180°) of the cylinder becomes low and attains minimum.

View full abstract

The reduction of drag of a circular disk of length/diameter ratio c/d=0.3 to 0.55 in axial flow was investigated. The drag coefficient of a circular disk with blunt shape of c/d=0.05 is about. 1.17 in the range Re=10⁴ to 10⁶. In the cases of c/d = 0.5 and 1.0, the drag coefficients are 1.13 and 0.91, respectively. To control the flow field around the circular disk, the side wall is cut to a spherical surface. Namely, the circular disk consists of two circular plane surfaces and a spherecal surface. The front and rear corners of the spherical surface located at 53°and 90°-120°from the front stagnation angle on a sphere, respectively. The surface oil-flow patterns indicated that the shear layer separated from the front corner reattaches on a spherical surface of the disk, and then the transition in the boundary layer and the turbulent separation occur. As the results, the wake width visualized by a smoke decreases drastically and the base pressure coefficient rises. The drag coefficient obtained by integrating the pressure distribution around the three circular disks with a spherical surface decreases by about a quarter of those of the three standard circular disks.

View full abstract

An experimental investigation has been made for a turbulent boundary layer near the trailing edge on a long flat plate. The flow was controlled by an additional splitter plate fitted to the trailing edge along the wake center line. The length of the splitter plate, l, was varied from a half, to five times the trailing edge thickness, h. Measurements of base pressure behind the trailing edge and of mean velocity and pressure distribution in the turbulent boundary layer on the flat plate were made under the freestream zero-pressure gradient. The absolute value of base pressure coefficient of the long flat plate was considerably smaller than that of the short flat plate (bluff body) without splitter plate (l/h=0). A significant increase in the normalized base pressure coefficient (about 50% as same as short one) was achieved with the splitter plate (l/h≥1) fitted to the long flat plate. Within an inner layer in the turbulent boundary layer near the trailing edge, mean velocity increased than that in the upstream position, especially in the case without splitter plate. With splitter plate, however, the base pressure rise made mean velocity distribution come closer to that of fully developed turbulent boundary layer. In the case of l/h≥1, the affected range of acceleration decreased from 10 h upstream of the trailing edge without splitter plate to 5 h upstream.

View full abstract

Investigation of a new high-speed zero-emission transportation “Aerotrain” is proceeded mainly in Tohoku University and University of Miyazaki. Since the Aerotrain utilizes the ground effect, researches of the aerofoil section which can efficiently harness the ground effect are important. The Aerotrain moves along U-shaped guide way which has a ground and two side walls, so a many viscous interference occur between the Aerotrain and the walls. The flow through primary and secondary aerofoils is analyzed to prevent the boundary layer separation for the improvement of the aerodynamic characteristics at low speed near the ground. A small secondary aerofoil is equipped above the trailing edge of the primary aerofoil to increase lift at takeoff and landing. The most suitable location of secondary aerofoil is investigated through the boundary layer approximation analysis and the experiment using a towing wind tunnel. We verify the effect of secondary aerofoil by experiment, confirm the accuracy of the analysis by comparing with experimental results, and try to find the most suitable aerofoil section for the Aerotrain.

View full abstract

When a plane jet exhibits flapping motion, two distinctive peaks, f₁ and f₂, appear in the spectrum of fluctuating velocity. The mean frequency is half of the frequency, f_v, of vortices shedding from a two-dimensional body that is installed in a jet nozzle contraction. Jet mixing has been enhanced by the flapping motion with a low frequency, f₃=f₁-f₂, which is considered to be caused by an inharmonic excitation. Vortical structures in the flapping jet are investigated by using flow visualization, image correlation and laser Doppler velocimetry. Vortices are seen to roll up asymmetrically at the jet nozzle exit, at a double frequency of that of the vortex shedding from the cylinder. Two types of vortices are formed by coalescence between these vortices and the vortices shedding from the cylinder. The periodicities of these two types of vortices are found to be equal. It is noted that vortices with f₁ or f₂ do not exist in the flow. The f₁ and f₂ are caused by the modulation of amplitude of the coalesced vortical structures with half of the flapping frequency, f₃.

View full abstract

In this study, unsteady cavitation and cavitation instabilities are numericaly analyzed around cascades in water/air working fluid. The numerical method employing a locally homogeneous model of compressible gas-liquid two-phase medium is applied, because it permits the entire flowfield inside and outside the cavity to be simulated easily through only one system of governing equations. Three types of.cavitation instabilities, which are similar to “cavitation surge”, “forward rotating cavitation” and “rotating-stall cavitation”, are represented numerically in flat-plate cascade under three-blade cyclic condition. Additionally, flowfields around flat-plate cascade with slit on blade are analyzed in the conditions where the three types of cavitation instabilities arise in the cascade without slit. The effect of the slit on suppression of cavitation instabilities is confirmed from the present numerical results.

View full abstract

A theory of a three-dimensional supercavitating cascade in spanwise shear flow between two parallel planes is presented. An equation of motion with respect to disturbance pressure is transformed into two problems by separation of variables. One of them is that of spanwise direction and the other is that of sectional plane of a hydrofoil. For the problem of the spanwise direction, an existing analytical solution is adopted. The problem of sectional plane of the hydrofoil is reduced to a simultaneous integral equation with respect to singularity distributions and is solved by using the analytical method previously proposed by the first author for the three-dimensional cascade without cavitation. These two kinds of solutions corresponding to individual eigenvalues are combined linearly, and expressions of the cascade and cavity characteristics are obtained. Numerical calculations are carried out to clarify the characteristics.

View full abstract

Evolution of chemically reacting flow has been studied experimentally using Hele-Shaw Cell.Neutralizaftion reaction of HCl+NaOH with different pH indicators is adapted for the chemical reaction. Low density liquid (HCl) and high density liquid (NaOH with different pH indicator) pour into upper layer in the cell and lower layer in the cell, respectively. In HCl+NaOH with 10% Bromothymol Blue (BTB) system, although some small proturbulences form and disappear repeatedly at a reaction front, flow patterns are stable and the front moves almost parallel to a horizontal line with time elapses. The higher initial concentration of reactants hastens the chemical reaction. On the other hand, in HCl+NaOH with 10% phenolphthalein (PP) system, unstable flow patterns are visualized, the chemical reaction is higher than that in the HCl+NaOH with BTB system. By the chemical reaction, temperature increases locally in the cell and density of product is smaller than that of HCl. It seems that the local temperature difference and density difference bring on the unstable flow patterns.

View full abstract

The two-phase fluid loop is very promising in thermal performance as one of the thermal control systems for satellites since it can transport higher quality of heat per unit surface area compared with the single-phase flow loop. Heat transfer performance of the evaporator with communication wick was investigated as one of the important constitutive components of the two-phase fluid loop. First, heat transfer characteristics of the evaporator with wick were analyzed. An evaporator with communication wick was manufactured on the basis of the analytical evaluation and the experiments of the heat transfer in the two-phase fluid loop were carried out using ethanol as the working fluid. As a result, it was found that the overall heat transfer coefficient showed the value of maximum 20 kW/m² for ethanol in experiment and had the higher value in the case of the working fluid, ammonia, in the analysis.

View full abstract

In this paper the thermal convection field and its resonance phenomena in a rectangular cavity with heat-flux vibration were numerically examined and the results were compared with the ones of square cavity. In the case of the aspect ratio α=5 and the Prandtl number Pr=0.71, for a range of the Rayleigh number Ra the critical angular velocity ω_c, at which the relative amplitude of midplane Nusselt number α_m has a local maximum, agrees very well with the resonance angular velocity of internal gravity wave ω_γ, estimated by the theoretical equation proposed by Thorpe, as observed for α=1. However α_m has two local maxima for larger Ra, which is peculiar to the case of larger α_m The time variation of components of fluctuating midplane Nusselt number shows that the phase at the maximum value of α_m agrees well with that of the fluctuating component of velocity for the first resonance angular velocity ω_γ. For the other angular velocity ω_γ2 the phase of α_m agrees with that of the fluctuating component of temperature. Moreover, we found that the boundary angular velocity ω₀ between the first two ω regions among five regions, which classify the thermal convection fields against ω, can be expressed by a function of a, Rα and Pr and that α_mis independent of α and Rα for relatively wide range of ω/ω₀.

View full abstract

In order to clarify the effects of liquid viscosity on inception of disturbance waves and droplets, structure of gas-liquid interface and flow rates of droplets entrained with gas were measured simultaneously at five kinds of kinematic viscosities, 1.0, 3.2, 9.9, 30 and 70 mm²/s. Time-spatial traces of liquid film thickness measured by five holdup probes were presented and it was found that the inception of disturbance waves occurred at Re_l= 200 or at the non-dimensional liquid film thickness of t_fm⁺=6.5. Furthermore, entrainment ratios were measured and it was clarified that droplets occurred before the inception of disturbance waves with increasing liquid kinematic viscosity and the critical condition of the droplet inception was j_l=0.02-0.03 m/s. The obtained results were compared with the published criteria on the inception of droplets and the previous correlations were found unsatisfactory. Considering the balance between shear force and surface tention acting on the wave crests, a new criterion of droplet inception was also proposed.

View full abstract

In the present study, a sooting flame was formed in a reduced-pressure combustion chamber using the vitiated coflow burner with a premixed flame of methane and oxygen for the surrounding flame and premixed flame of toluene, methane and oxygen for the center flame. Collected soot samples were analyzed using the high performance liquid chromatography (HPLC) and a gas chromatography to determine the fullerene content of soot and the species of PAHs under various conditions. Effects of pressure and equivalence ratio to the formation of the PAHs and fullerenes and the branching mechanism between fullerene and soot were discussed. When the fullerene content in soot samples was relatively high, the fractions of anthracene and benzo [a] anthracene were relatively high. When the fullerene content in soot samples was relatively low, the fractions of phenanthrene, flouranthene and pyrene were relatively high. Under the same pressure condition, there was an optimal flow rate of toluene in the center flame for formation of C₆₀ and C₇₀ of which range became wider with the decrease of pressure condition.

View full abstract

As a means for a stable premixed combustion, there is a so-called cyclone combustor, which consists of a cylindrical chamber and fuel nozzles installed tangentially on the side wall. In this combustor an extremely stable flame can be obtained in the swirl flow, formed along the inner wall of the combustor. The authors utilized this combustor as a flame holder, to burn a high velocity jet flowing axially in the central part, and named this combustor a cyclone-jet combustor. In general, it is difficult for the low NO_x combustion to occur in spray combustion. In the present study, experiments on the spray combustion by the cyclone-jet combustor were carried out, and comparing the co-axial fuel spray and fuel gas jet flames. Kerosene and propane were used for fuel. With the air flow rate changed, the effect of the premixing on the NO_x formation was examined. It was shown that the increase of the air flow rate reduces the NO_x emission, and the shape of the spray flames is almost the same as the gas flames. The results suggest that the premixing after the process of the injection of fuel becomes very effective for NO_x reduction in spray combustion.

View full abstract

A new low NO_x combustor for kerosene-fueled micro-gasturbine was proposed, and the combustion characteristics of prototype combustor were investigated. In order to enhance the recirculation in a primary combustion zone, a swirler was set between the primary and secondary combustion zones. Combustion air was introduced through the swirler and forced to flow upward to the combustor bottom, from where the fuel spray was supplied through a nozzle. To achieve high combustion stability and low emission in wide fuel-air ratio, the optimum configuration of the primary combustion zone were investigated. The optimum one was found out by measuring the fundamental combustion characteristics such as lean combustion limit, flame luminosity, exhaust gas composition and combustion gas temperature.

View full abstract

The chemiluminescence spectra were obtained sequentially during the Homogeneous Charge Compression Ignition (HCCI) combustion. At the same time, the combustion processes were photographed using a high-speed camera with image inensifier or a color high-speed camera. The image of the cool flame was captured, and the Emeleus cool flame band was observed. Then blue flame became evident. During the main heat release duration, the hot flame emission spectrum was observed where the distinct OH emission appeared, and it was superimposed on the carbon monoxide continuum. After the main heat release, the broad peak of the spectrum was shifted to the longer wavelength side, and the images of the flames change to red, especially at higher equivalence ratios. This red coloration must be originated from the H₂O vibration-rotation bands. The chemiluminescence spectra and the color high-speed images of combustion initiated by Pulsed Flame Jet (PFJ) were also obtained. In PFJ, the emission spectrum obtained was composed of the distinct OH emission and CH emission superimposed on the carbon monoxide continuum.

View full abstract

Since the engine combustion is affected by a variety of parameter such as well as fuel characteristics, it is very difficult to find out the clear effect of fuel property. In order to clarify the change of emission characteristic caused by fuel change, we propose a fundamental flame research method using a small pool-wick flame system. By using this system, PAHs (polycyclic aromatic hydrocarbons) formation mechanism in a flame was investigated as a fundamental study of diesel combustion. Characteristics of PAHs formation in a laminar diffusion flame were measured by a laser induced fluorescence (LIF) method. It was compared with PM charactersistics measured by a laser induced incandescence (LII) method. The effect of the lubricating oil contamination on the PAHs distributions in the flame was discussed. As the result, it was found that the PAHs concentration in the flame increased with increasing the lubricating oil contamination and it caused the increase of PM formation in the flame.

View full abstract

In a constant volume combustion chamber, a study of investigating for relationship between NO_xemissions and non-homogeneity distribution of methane-air mixture was conducted. Non-homogeneity distribution was generated by varying the period after fuel injection and movement of perforated plate. Non-homogeneity distribution was measured by the acetone planar LIF technique and the heterogeneity index was determined. The exeriments were performed under high pressure and temperature conditions in the chamber. The effects of heterogeneity index and initial condition in the chamber on NO_x and conbustion efficiency were studied in detail. The trend of NO_x and combustion efficiency was varied in different heterogeneity mixture at the same excess air ratio.

View full abstract

In order to investigate early soot formation process in diesel combustion, spectral analysis and optical thermometry of early soot formation region in a transient spray flame under diesel-like conditions (Pg=2.8 MPa, Tg=620-820 K) was attempted via laser-induced fluorescence (LIF) from pyrene (C₁₆H₁₀) doped in fuel. Pyrene is known to exhibit a temperature dependent variation of LIF spectrum, the ratio of S 2/S 1, fluorescence yields from the lowest excited singlet state S 1 and the second excited singlet state S 2, depends strongly on temperature. In the present study, pyrene was doped (1% wt) in a model diesel fuel (0-solvent) and the variation of LIF spectra from the pyrene in the spray flame in a rapid compression machine were examined at different ambient temperatures, oxygen concentrations, measurement positions and timings after start of fuel injection. Nd : YAG-pumped dye laser pulse at 309 nm (5 mJ) was used as the excitation source and the LIF spectra were measured by a spectrometer. The pyrene LIF spectra measured in non-combusting evaporating spray in nitrogen atmosphere showed increase in the S 2/S 1 ratio of LIF intensity and shift of the peak wavelength of S 1 band towards red with increasing ambient temperature. The pyrene LIF spectra measured in spray flame between start of injection and ignition showed increase in the S2/ S1 ratio of LIF intensity and shift of the peak wavelength of S 1 band towards red as the time progresses after start of injection, showing that the temperature variation of the gas mixture up to ignition can be captured by this technique. The LIF spectra measured in spray flame after ignition showed increase in emission intensity in longer wavelength region due to LIF from combustion products and the S 2/S 1 ratio of LIF intensity decreased. The pyrene LIF spectra measured in spray flame at 1.0 ms after start of injection at different ambient oxygen concentrations (5 to 21%) showed increase in the S 2/S 1 ratio of LIF intensity and shift of the peak wavelength of S 1 band towards red with increasing oxygen concentration, showing that the temperature difference of the gas mixture caused by progress of reaction at different oxygen concentrations can be captured by this technique.

View full abstract

The dependence of ultra-high EGR and low oxygen diesel combustion on fuel properties was investigated with a single cylinder naturally aspirated 1.0 L DI diesel engine. Decreasing cetane number in fuels significantly reduces the smoke emission due to an extension in the ignition delay and the subsequent improvement in mixture formation. Smokeless combustion, ultra-low NO_x and efficient operation range for EGR and for IMEP are remarkably extended with decreasing the cetane number. Change in fuel distillation temperature does not result in significant differences in the smoke emission and thermal efficiency at ultra-high EGR operation, and smokeless operation is established even with higher distillation temperature fuels as long as the cetane number is sufficiently low.

View full abstract