A computational method has been proposed for viscous incompressible flows accompanied by free-surface oscillation. In order to improve the numerical accuracy for internal flows, QSI (quintic spline interpolation) scheme is applied to convection terms. The free-surface profiles are represented by curvilinear coordinates generated on the basis of ALE formulation, in which the governing equations are discretized on the collocated grid system. The velocity-pressure correction is performed with C-HSMAC method, which is effective to preserve fluid mass during the unsteady computation with free-surface deformation. The computational method was applied to small amplitude standing waves, non-linear waves and lid-driven free-surface flows in a cavity. As a result, it has been shown that the computational method has sufficient numerical accuracy for free-surface flows.

View full abstract

A finite difference method for the simulation using Cartesian grid is developed. As the body boundary is not necessarily located on the grid points, finite difference formulation near the boundary is critically important. The distance from the adjacent stencils to the body boundary is integrated into the finite difference formulation to satisfy the body boundary conditions. From the stability analysis, it is proved that the mid-term in the dissipation term should be treated implicitly to avoid the severe diffusion number condition. The artificial dissipation term based on the reconstruction that uses a virtual center point gives the most accurate solution near the body boundary. As for the pressure boundary condition, the internal pressure is defined by the least square method to satisfy all the differential calculus along the normal vector to the body boundary should be zero as much as possible around the stencil.

View full abstract

Diatomic molecular collisions are simulated by the classical trajectory calculation (CTC) technique with the flexible model using the Lennard-Jones potential and the Murrell-Sorbie potential. The trajectory calculation using the flexible model is combined with the direct simulation Monte Carlo (DSMC) method and is examined by simulating rotational relaxation of nitrogen in a homogeneous space. The CTC-DSMC method with the flexible model simulates reasonably the rotational relaxation for the wide range of temperature. The CTC-DSMC method with flexible model is applied to the simulation of normal shock wave structures of nitrogen for upstream Mach numbers of 1.71, 7.0, and 12.9. It is confirmed that the calculation results, such as the number density and rotational temperature profiles and rotational energy distribution across shock waves are in reasonable agreement with the experimental results of Robben and Talbot.

View full abstract

A bubble or droplet can be driven in a microchannel due to the thermocapillary force controlled by microheaters, which comes from the surface tension imbalance. Though practical optical switches based on this principle have been developed by micromachining technique, there has been no contribution of fluid dynamics both experimentally and analytically. In order to treat gas-liquid interface accurately in a microchannel, it is necessary to improve the models of surface tension and contact angle. A new numerical scheme is developed based on C-CUP and Level-Set function method with CSF model and is able to calculate the surface tension distribution precisely. It is the first time to simulate the thermally-driven bubble in a microchannel and the obtained results show good agreements with experimental results qualitatively. Comparisons with the conventional numerical schemes and fundamental mechanisms of the thermally-driven interfaces are also discussed.

View full abstract

When an air-column in an axi-symmetric closed tube with constant cross-section is driven by a piston vibrating at resonant frequency at one end, finite amplitude standing wave is induced in the tube. The increase of oscillation amplitude in the tube brings about the nonlinear phenomena of wave motion, such as mean pressure distribution along the tube axis, acoustic streaming, and thermoacoustic effect. The large amplitude pressure fluctuation and the thermoacoustic effect generated by finite amplitude standing wave induced in the closed tube are principal physical phenomena in the engineering applications of the finite amplitude wave motion to acoustic compressor and thermoacoustic refrigerator. In order to develop these practical systems, it is essential to realize large amplitude oscillation in the tube. However, the increase of oscillation amplitude is limited by the acoustic saturation caused by nonlinear wave distortion which steepens wave front to shock wave. In order to avoid the acoustic saturation, axi-symmetric tubes with cross-sectional area contraction toward the closed end are used. Basic characteristics of wave motion in the tubes with area change have been discussed with linear acoustic theory. Finite amplitude wave motions generated in different shape tubes have been analyzed numerically and compared with experimental results. Fundamental knowledge and estimation for the development of the acoustic compressor have been given from the results obtained.

View full abstract

The purpose of the present paper is to compare the characteristics of the weak Mach reflection (MR) and the so-called von Neumann reflection (vNR). A series of experiments was performed to investigate both reflections. Attention is focused on the flow properties around the triple point. By measuring the location of the triple point and the angle between the incident and reflected shocks, we estimated the approximate flow properties around the triple point. The results show that, while the triple point moves along an almost straight trajectory through a wedge tip, the angles of incidence and reflection vary as the incident shock propagates. As a result, the well-known self-similarity law does not hold for both reflections. For the von Neumann reflection, the present experimental data vary almost along the trivial solution curve of the von Neumann's three-shock theory, and an asymptotic state seems to exist on this curve. In contrast, for weak Mach reflections, these data do not lie on the trivial solution curve but reveal the so-called von Neumann paradox.

View full abstract

Controlling mechanism of the crossflow dominated transition process in the three-dimensional boundary layer using a small sized roughness is investigated. The stationary crossflow vortices are excited by placing the first roughness elements for the wavelength of λ_n = 12.5 mm in the spanwise direction, which is equal to the most amplified wavelength under the natural condition, on a swept flat-plate model. The second roughness elements are located three different points between the stationary crossflow vortices. We found that the characteristic asymmetric structure of the stationary crossflow vortices becomes symmetric so that the high shear regions in the vortices are diminished when we located the second roughness at - 1/3 of λ_n with respect to the first roughness. As a result, generation of the high frequency secondary instability is overdue and the delay of turbulent transition has been attained.

View full abstract

Developing air-water bubble flows in a vertical pipe were calculated using the combination of a multi-fluid model and an interfacial area concentration (IAC) transport equation. Calculations were conducted by using (a) a standard two-fluid model, (b) a two-fluid model and a one-group IAC equation, (c) a two-fluid model and a two-group IAC equation and (d) a three-fluid model and a two-group IAC equation. As a result of comparisons between measured and calculated void fractions, we confirmed that (1) the use of IAC equations brings about the reconstruction or reconsideration of interfacial drag models, (2) the two-group IAC equation requires initial values for a large bubble group even for the condition of no large bubbles, (3) the combination of the two-group IAC equation and two-fluid model cannot give good predictions for flows consisting of large and small bubbles, and (4) a multi-fluid model and a reconstructed drag force model are indispensable to obtain good predictions for bubbly to slug flow transition.

View full abstract

To obtain the time-response and the effect of temperature of the pressure sensitive paints(PSPs), the luminescence intensity of the PSPs is measured in the pressure and wall temperature range of 1∼101.3 kPa and 263∼303 K, respectively. The PSPs studied in the present works are PtOEP, PtTFPP, H₂TFPP, H₂TCPP, H₂TSPP, and H₂TTMAPP; they are painted on a commercially-available porous silica thin-layer chromatography (TLC) plate. The effect of the plate temperature on the pressure measurements is examined, and it is found that H₂TFPP has the smallest temperature sensitivity of 3.5 kPa/K among the paints tested. The time-response of the paints is also investigated by using a shock tube. The change in luminescence caused by the sudden pressure rise due to an incident shock wave is detected by a photomultiplier. Results indicate that the response time is on the order of 20∼40 microseconds. In addition, the contact surface in the shock tube flow provides a method of checking the effect of the sudden change in gas temperature, and it is found that the luminescence intensity is insensitive to this effect.

View full abstract

Time response of pressure-sensitive luminescent coatings has been investigated theoretically and experimentally. The present coating consists of a thin anodized aluminum layer and luminophores. The layer is formed onto the surface of pure aluminum by an electro-chemical process. The luminophores are adsorbed onto the surface of the layer via chemical and physical adsorption. A method of making this coating is described in detail. The theoretical analysis shows that the effective diffusion coefficient for oxygen permeation in the anodized aluminum layer is up to 1 × 10^-5 m²/s. This implies that the present coating should have the response time of the order of microseconds. For three kinds of luminophores, two porphyrin compounds and a ruthenium(II) complex, the response to a step change in pressure was studied using a pressure jump apparatus and a shock tube. It has been found that the response time of coating with tris(4, 7-diphenylphenanthroline)ruthenium(II) ([Ru(dpp)₃]²+) is longer than 20 μs, and depends on the thickness of anodized aluminum layer. On the other hand, tetrakis-(4-carboxyphenyl) porphyrin (TCPP) coating has the time response less than 10 μs, which is independent of the thickness of layer.

View full abstract

The fluctuating and time-averaged aerodynamic forces acting on two circular cylinders and characteristics of flow over the cylinders in tandem arrangement were investigated experimentally at a Reynolds number of 6.5 × 10⁴. The position of the downstream cylinder was varied systematically up to the spacing of 9 diameters and time-averaged drag, fluctuating pressure, fluctuating lift and drag forces acting on the cylinders were measured. Two peaks in each of the distribution of the fluctuating lift and drag forces acting on the downstream cylinder were observed when the cylinders were placed before the critical spacing. The value of the fluctuating lift and drag forces at the larger peak that occurred at a spacing ratio of 1.4 was about 2 and 2.8 times, respectively that of a single cylinder. The magnitude of the fluctuating lift and drag forces acting on the downstream cylinder was related very sensitively with the reattachment position of the separated shear layers from the upstream cylinder onto the downstream one and the fluctuating lift and drag forces became maximum when the reattachment position of the shear layers reached minimum and vice versa.

View full abstract

Flows around a decelerated circular cylinder are studied with BFC-MAC scheme for a moving frame of reference. The main results are given for Reynolds number lower than Re=36.5 as in the previous experimental measurements. Typical behaviors of wake length are predicted at a range of deceleration as in the experiment. That is, histories of wake-length: S, if they are plotted against instantaneous Reynolds number, are almost similar for relatively large deceleration α. Such 'Similarity' with respect to α is attempted to be interpreted with a model equation based on approximate estimation of arrival time of the wake-end. In contrast at small deceleration the transient variation of wake-length depends upon α. The mechanisms are discussed such as the fact that the smaller deceleration parameters, the larger delay in elongation of wake length. For large deceleration the proposed model equation, which utilizes the BFC-MAC results on the rear symmetry line at an arbitrary couple of time-steps, successfully illuatrates wake-length history versus instantaneous Reynolds number from start to termination of deceleraton.

View full abstract

The two types of longitudinal vortices, the trailing vortex and the necklace vortex, shed periodically near the intersection part of the criss-cross circular cylinders located in uniform flow. The influence of the oscillation of upstream cylinder on longitudinal vortex excitation caused by their vortices was investigated. In the case where the trailing vortex is shedding, the lock-in phenomenon occurs on the wide condition for the free stream velocity and the frequency and the amplitude of the upstream cylinder oscillation, compared with the lock-in condition of Karman vortex. On the other hand, the lock-in region of the necklace vortex is narrower than that of Karman vortex. In either case, the phase difference between the oscillation of the upstream cylinder and the vortex shedding decreases with increasing non-dimensional oscillation frequency.

View full abstract

The self-sustained oscillation of the shear layer occurs in a round jet impinging on a ring placed coaxially. In the case of the finite length pipe instead of the ring, the oscillating frequency of the shear layer is locked in to the resonant frequency of the pipe in frequency, when both the frequencies approach each other. This study aims at getting a clearer understanding of the lock-in phenomenon in round jet-body system. The experiments are practiced for the four cases. In the two cases a resonator like a pipe is located downsteram to the nozzle exit, in the other two cases a resonator consists of a part of the nozzle with a ring downstream. As a result of the experiments, it has become clear that the lock-in phenomenon takes place in all the cases, but the intensities of the lock-in and the generating resonance are different, depending on whether a resonator is located downstream of the nozzle exit or at the nozzle.

View full abstract

The effects of particle density on turbulence statistics, such as the preferential concentration and the particle dispersion, are numerically investigated for various Stokes numbers. The small particle motions in isotropic homogeneous turbulent flow are simulated by the one-way coupling method. In the present method, the fluid and particle phases are solved in the Eulerian and Lagrangian ways, respectively. The preferential concentration for zero-density particles, (i.e. corresponding to the contaminated bubble limit), is much stronger than that for the heavier particles. The root-mean-square velocity of the particle "υ rms" versus the Stokes number is analyzed through the comparison of the Tchen's theory (Hinze, 1975). The υ rms profile for the heavier particles shows good agreement with that of the Tchen's theory, while that for the bubbles shows considerable difference from the theory due to the strong preferential concentration. The dispersion coefficient of particle "D_p" is also investigated. At the Stokes numbers, such that the preferential concentration occurs, D_p for the heavier particles is larger than the dispersion coefficient of fluid "D_f", while D_p for bubbles is smaller than D_f. The difference between D_p and D_f for the bubbles is much larger than that for the heavier particles.

View full abstract

Due to the translational motion of a bubble, drop or particle, the pressure averaged over its surface becomes less than the bulk one. In the present study, a dimensionless number "C_TDP", which corresponds to the dimensionless pressure reduction due to the translational motion scaled by the fluid density and translational velocity, is theoretically investigated. The particle Reynolds number "Re" is assumed to be sufficiently small but non-zero and the method of matched asymptotic expansions is used to evaluate C_TDP as far as the term of order Re² log Re. The result shows that the leading order of C_TDP is Re⁰. The O(Re² log Re) component of C_TDP is affected by the outer field of the expansion, while the O(Re⁰) and O(Re¹) compenents are not.

View full abstract

High-speed pumps, such as turbopumps for rocket engines, are usually operated with cavitation. The cavitation often becomes unsteady and generates the unsteady fluid forces on the blades. For the development of turbopumps with higher reliability, it is important to clarify the unsteady forces and pressure distributions on the blade. In the present study, the fluid forces and unsteady pressure distributions caused by flow rate fluctuation and inlet pressure fluctuation were calculated under the condition with cavitation. The effects of cavity length and the frequency of fluctuation on the unsteady forces and pressure distributions are discussed for lower and higher solidity cascades.

View full abstract

An analysis of cavitation in flat plate cascades under the interaction with sinusoidal gust flow was carried out by a singularity method based on the closed cavity model allowing the cavity length to change freely. The fluid forces and pressure distributions on the blade were calculated. The effects of cavity length and the frequency of the gust flow on the unsteady forces and pressure distributions are discussed for lower and higher solidity cascades. The results are compared with experiment on 4-bladed inducer under the interaction with inlet distortion.

View full abstract

In the present study, a higher order rotating cavitation predicted by the stability analysis was identified through the measurements of inlet pressure fluctuations and blade stress fluctuations. The propagation speed ratio of the higher order rotating cavitation is about 5, and the amplitude of the blade stress fluctuation at the root of leading edge caused by the rotating cavitation is the same level as the conventional rotating cavitation. The higher order rotating cavitation occurred at higher cavitation number compared with that of conventional rotating cavitation. A higher order cavitation surge was also observed at the transition point from the conventional to the higher order rotating cavitation. It was clarified that these cavitation instabilities can be suppressed by modifying the shape of the inlet casing.

View full abstract

We have developed a new Stokesian dynamics (SD) method for non-dilute colloidal dispersions, which enables us to reduce drastically the computation time. To verify the validity of the present method, which is called "the cluster-based SD method", the simulations of a ferromagnetic colloidal dispersion have been carried out for a simple shear flow. The correlation function and the viscosity have been evaluated in order to compare the results obtained by the present method with those obtained by the ordinary SD method and the method of ignoring hydrodynamic interactions between particles. The results obtained here are summarized as follows. The transient properties from an initial state obtained by the present method agree well with those by the ordinary method, even if the radius r_clstr, which defines the cluster formation, is taken as a small value such as r_clstr=1.2d (d is the particle diameter). Also, concerning the equilibrium properties such as the pair correlation function and the viscosity, the present cluster-based method gives a satisfactory agreement with the ordinary SD method. Furthermore, the cluster-based method drastically reduces the computation time by about one fourteenth to one seventieth compared with the ordinary method. It is clear from these results that the cluster-based SD method is significantly superior to the ordinary SD method for ferromagnetic colloidal dispersions for which a large model system such as N=1000 or 10000 is indispensable in simulations.

View full abstract

Many researches of drag reduction caused by cationic surfactants have been reported in recent years. However, there are few papers about the application of drag reduction for practical air conditioning systems. In this study, the effects of corrosion inhibitors on the drag reduction caused by a cationic surfactant, Ethoquad O/12, were experimentally investigated. Suitable corrosion inhibitors were proposed and significant change of rheological properties of surfactant solutions with the corrosion inhibitors was also indicated. Such drag-reducing additive was used for a practical air conditioning system, and the energy saving rates was experimentally obtained. Other ten examples of drag reduction adopted for practical systems were also presented in this paper.

View full abstract

This paper is related to the deign optimization of axial flow hydraulic turbine runner. An improved Q3D inverse method has been proposed from the system point of view, and a set of rotational flow governing equations as well as an equation for blade geometry design have been derived. The computation domain is firstly taken from the inlet of the guide vane to the far outlet of the runner blade, and flows in different regions are solved simultaneously. Therefore, the influence of wicket gate parameters on runner blade design is considered and the difficulty of defining the flow condition at the runner blade inlet is overcomed. Because a pre-computation of initial blade on the S₂m surface is newly adopted, the iterations of S₁ and S₂m surfaces has been greatly reduced. Experimental results have validated the present model and a direct flow analysis has demonstrated the reliability of the inverse computation. Numerical investigations show that the effect of wicket gate parameters should be considered also in runner design.

View full abstract

This paper is concerned on the design optimization of axial flow hydraulic turbine runner blade geometry, and a new comprehensive performance Optimization procedure is presented by combining a multivariable multi-objective constrained optimization model with a Q3D inverse computation and a performance prediction procedure. The total hydraulic loss and the cavitation coefficient are taken as optimization objectives and a comprehensive objective function is defined using weight factors. Optimization variables are taken from parameters describing the blade bound circulation distribution and positions of the blade leading and trailing edges in the meridional flow passage. The optimization procedure has been applied to the design optimization of a Kaplan runner. Numerical results show that the performance of the designed runner is successfully improved through optimization computations. With the multi-objective optimization model, it is possible to control the performance of designed runner by adjusting the value of weight factors defining the comprehensive objective function.

View full abstract

The developed rotary heat exchanger has three principal functions as follows, a blower, a heat exchanger and a pump. Structure is similar to cross flow fan. The blades also work as heat transfer surfaces. Further more, this rotor is equipped many circular tubes, liquid collection and distribution tanks to flow the cooling liquid. So previous studies about cross flow fans are not able to apply to this type of rotor directly. We prepared various samples which have different parameters in rotor diameter, length, diameter ratio of inner to outer and fin pitches. Throughout series of studies, we found the relation between the airflow rate and various parameters that mentioned before. A stationary guide impeller attached to the outlet side of rotor and a rotary impeller attached to the inlet side are necessary to support the pumping action. And we indicate the relation between the water flow rate and the impeller's shapes.

View full abstract

Rotary-heat exchanger which have three functions as a blower, a heat exchanger and a pump have been developed on the basis of a cross flow fan. In the first paper, the blower and pump performances of the rotary-heat exchanger were clarified. In the present paper, we report the heat transfer characteristics of it as a heater and a cooler. The experiments as a heater have been done using with several kinds of rotors that differ in shape. As a consequence, the experimental relation between the heat transfer coefficients and the parameters of the rotary-heat exchanger was obtained. And in the experiments as a cooler, it is clarified that the heat transfer coefficient is higher than that of a heater. In particular, condensed water droplets on the blades of rotor which are heat transfer surfaces are rapidly and continually blown off by the centrifugal force.

View full abstract

Natural convective flows over upward-facing horizontal plates were investigated experimentally. The main concerns were focused on the effects of Prandtl numbers on the flow and heat transfer characteristics, in particular, of the laminar and transitional regions. Air and ethylene glycol were utilized as the test fluid and the test plates were heated with uniform heat flux. The flow and temperature fields over the plates were visualized with smoke and liquid crystal sheet to investigate the turbulent transition process. The result showed that three-dimensional separation of main flow becomes a trigger to turbulent transition and that the onset of the separation is described with the modified Rayleigh numbers as (3-4) × 10⁶ for both air and ethylene glycol. The local heat transfer coefficients of the plates were also measured and the results were compared with previous results for water. The comparison showed that the coefficients become higher with the Prandtl numbers. Based on these results, empirical heat transfer correlations that account for the Prandtl number effect were proposed.

View full abstract

Experimental study has been conducted on local heat/mass transfer and flow characteristics in rectangular cross-sectioned two-pass channels with a sharp turn having an inclined divider wall. Detailed distributions of local Sherwood numbers are presented that are obtained on all the walls of the channels by the naphthalene sublimation method, and the influence of the inclination angle of the divider wall on the local heat/mass transfer are examined. Then the flow characteristics in the channels, such as flow separation and reattachment around the turn section, secondary flow velocities in and after the turn, are described based on the time-mean flow vectors obtained by using the PIV. The correlation between the local heat/mass transfer and flow structure is discussed.

View full abstract

The study presents a numerical analysis on the thermal behavior of frozen-melting boundaries produced in the permafrost area subjected to the periodic changes of surface temperature. The simplified analytical model is introduced to postulate a single component, instead of water-contented soil, for taking a higher priority to reveal the thermal structure in the underground and the heat transfer mechanism across the surface as well as the frozen-melting boundaries. The analytical results show the temperature profiles produced in the active layer to disclose the drastic changes for the coalescence of two adjacent frozen solid layers separated by the liquid and vice versa. The temperatures in the deep underground, maintained almost constant similar to the Stokes 2nd problem, are found to be dereased to rather lower compared with the mean temperature at the surface. The futher detailed discussions are executed on the heat transfer mechanism on the transient phenomena in the mid- and long-term perspectives of environmental effects and climate changes.

View full abstract

This paper has dealt with cold heat extraction characteristics from the gel latent-cold-heat storage-material suspension with surfactant to air bubbles. The gel latent-cold-heat storage-material consisted of n-paraffin as the core latent-cold-heat storage-material and water as a heat transfer medium. The relationship between outflow air temperature in latent-cold-heat release process and various parameters was examined experimentally. As a result, especially concentration of the gel latent material added to water exerted an influence on gas holdup and cold heat extraction characteristics to air bubbles. The non-dimensional correlation equations for the temperature effectiveness of latent-heat storage process were derived in terms of the ratio of water layer height to diameter of latent heat material, Reynolds number of air flow, Stefan number and modified Stefan number including air humidity.

View full abstract

Driving device for heat transfer enhancement in a fluid layer heated from above was developed. The device was composed of a coil spring made of shape memory alloy and a closed vessel that is flexible in axial direction telescopically. It shrinks and sinks under water when it is hot, and it expands and floats on water when it is cold. Therefore, the device was named "Negative Thermal Expansion Capsular Object" shortly NTE capsule. Two types of the NTE capsule were proposed. One is the inner spring type and the other is outer spring type. Their performance was proved experimentally. Their operating characteristics were compared to each other and their potentialities in heat transfer enhancement were discussed. As a result, it was proven that the heat transfer enhancement effect could desire NTE capsule of the outside spring type.

View full abstract

An increasing demand for high reliability of electronic equipment requires free convection cooling of electronic components on printed wired boards (PWBs). The electronic components on the PWBs are cooled with natural ventilation by air flow and the temperature rise of the ventilated air between PWBs must be accurately predicted. These subjects make the thermal design difficult for equipment designers, and the thermal optimization method using design knowledge and numerical analysis model before making proto-type are needed. In this paper, the thermal phenomenon in a model enclosure which had 16 parallel vertical smooth flat plates with heaters was studied experimentally. It was shown that the ventilated flow rate in one section was found to be affected by the heat loads in other sections. And the predicted temperature rise with Computational Fluid Dynamics model was in good agreement with the experimental results.

View full abstract

Purpose of this paper is numerically to make a resolution of interactions between hypersonic reacting flows and wedges. Gas utilized in this study is a stoichiometric hydrogen and oxygen gas mixture which composes 19 elementary reactions with 8 chemical species. A finite difference approximation used in the simulation consists of the MacCormack-TVD scheme and the semi-implicit scheme which treates implicitly chemical reaction terms. The flow Mach number and the half vertical angle of wedge as flow parameters were changed from M=5 to 10 and θ=15 to 45 degree, respectively. It was shown in the inviscid flow that an oblique shock wave front and an detonation wave front coexist on the wedge surface depending on a combination of the Mach number and the wedge angle. In the viscous flow model, only the detonation wave front was observed within the content of the calculation conditions.

View full abstract

This paper is concerned with a new approach that produces numerically a detonation with a triple-shock-waves structure. Hitherto, propagation behavior of gaseous detonations has been analyzed by solving Euler equations including chemical reactions under a combination of a solution of one-dimensional C-J detonation and some pairs of disturbances with different exothermicity as an initial condition. That is, addition of any disturbances such as exothermicity in the vicinity of the initial wave front was necessary as far as Euler equations were treated. However, in this paper, Navier-Stokes equations were solved by using only solution of the one-dimensional C-J detonation as an initial condition. The present method is more convenient than the conventional method. As a result, it was found that numerical production of a two-dimensional unsteady detonation was possible by utilizing the semi-implicit MacCormack-TVD scheme together with an adequate mesh size which is less than Δx=Δy=0.09 mm.

View full abstract

In this study, numerical simulations of the counterflow diffusion flame and the counterflow double flame with oxygen-enriched air were made by using the GRI chemical reaction mechanism (NO_x included) in order to elucidate the effects of oxygen concentration and equivalence ratio of mixture on the flame structure and NO_x formation. The following conclusions were reached: (1) In the case of counterflow diffusion flame, as the oxygen concentration increases, the width of high-temperature zone becomes larger, so that NO formation by the thermal NO mechanism increases, but that by the prompt NO mechanism decreases and becomes negative. (2) In the case of counterflow double flame, as the equivalence ratio becomes large, the NO formation by the prompt NO mechanism at the position of the fuel-rich premixed flame decreases, and that by the thermal NO mechanism at the position of the diffusion flame keeps approximately constant. (3) In the both flames, NO formation has a maximum at about Y_O2=0.8, and it decreases steeply above Y_O20.8, because the nitrogen concentration decreases.

View full abstract

This paper presents discussions on predicting turbulent flow in a simplified internal combustion engine geometry consisting of an intake port, valve and cylinder. The low-Reynolds-number linear k-ε, cubic nonlinear k-ε and k-ε-A₂ turbulence models are applied to the flow field computation and evaluated against experimental data. The results suggest that the cubic nonlinear eddy viscosity modelling of turbulence is very successful for predicting mass flow rates and in-cylinder velocity distributions.

View full abstract

This paper presents a new concept of hydrogen internal combustion Stirling engine and experimentally clarifies the effect of the internal combustion by comparing performances of a prototyped engine on internally heating and externally heating. The hydrogen combustion Stirling engine utilizes internal combustion of stoichiometric H₂ and O₂ mixture injected into the working fluid gas as thermal input, and the cyclic operation completes with the removal of the water from the engine after the condensation at the cooler. The prototyped engine substitutes a catalytic combustor for the conventional heater and H₂-O₂ mixture is injected at a constant flow rate between the regenerator and the cooler. The engine performance was evaluated measuring temporal change in pressure and temperature in the expansion and the compression space, thermal input and rejected heat. The internal heating performance showed almost the same characteristics as that of external heating except for the increase of expansion work due to the direct thermal input. The increase of expansion work improved the engine performance, particularly in high engine speed region. Furthermore, the steady premixed injection method showed a possibility to suppress easily the mixture strength in working gas.

View full abstract