A 3-D linear analysis of rotating stall is carried out to clarify the effects of spanwise distribution of cascade characteristics. An actuator disk model of annular cascade is employed. A linear spanwise distribution of the performance slope is considered to examine the effects of the threedimensionality in the cascade characteristics. Many characteristic solutions were found corresponding to various kinds of radial mode. It was found that the 0th order radial mode is more unstable than higher order modes. It was also found that uneven spanwise cascade characteristics distribution makes the 0th mode more unstable, and alters the 0th order mode so that the amplitude of the disturbance becomes larger at radial locations where the positive slope of the performance is larger.

View full abstract

This paper deals with the factors degrading magnetic fluid functions. We have two kinds of coagulation of magnetic solid particles in a magnetic fluid. One is formed with a magnetic field and the other without it. The latter coagulation is considered to be caused by the deterioration of surfactants, or the decrease of their thickness around magnetic solid particles and the residual magnetization in them. Such a coagulation has a tendency to make a magnetic fluid inhomogeneous and degrade its function. We have tried to find out the factors affecting on such a coagulation using a centrifuge method and as a result, shearing force of fluid flow, aging of a magnetic fluid and residual magnetization of soild particles were recognized as to be such factors.

View full abstract

Velocity fluctuations in the near wake of an elliptic disk are analyzed by the wavelet transform with Morlet wavelet. Modulations of the wavelet coefficient demonstrate two dominant periodic components in the velocity fluctuations and their amplitude modulations. The periodic components, which correspond to frequencies of vortex shedding in the major-and minor-axis planes, have low-frequency amplitude modulations. The modulus of the wavelet coefficient shows that the amplitude modulations have a central frequency of approximately 1/5 of the corresponding vortex shedding frequency. Moreover, the amplitude modulations are found to be out of phase in the major and minor-axis planes, while those in each plane are in phase at opposite positions in the plane.

View full abstract

Start-up flows of viscoelastic fluids in a two-dimensional 4 to 1 abrupt contraction channel is numerically studied and the transient behavior of viscoelastic fluids is discussed. The multi mode Leonov model is applied to describe the rheological properties of the viscoelastic fluid. The numerical simulation has been carried out for two Weissenberg numbers. The following results are obtained from the present analysis : The corner vortices develope with time and restrict the velocity gradient to relax the growth of the stress field. The circular secondary flow works as a stress relief mechanism. The overshoot of the stress growth is observed at the high Weissenberg number and is not done at the low Weissenberg number. This phenomenon is caused by the unsteady elongational property of the fluid.

View full abstract

Mechanism of transmission behavior of a pulse deformation on a viscoelastic fluid jet is investigated. The jet of the viscoelastic fluid falls down in air vertically due to gravity. The pulse deformation is formed on the jet by impact of collision with a swinging rod. The deformation transmits to upward on the jet and rebounds off the nozzle exit like wave motion on a string. The motion of the deformation is recorded by a video camera to evaluate its transmission velocity. The transmission velocity decreases with increasing flow rate and it demonstrates the wave transmitts with a velocity relative to the falling jet liquid. The relative transmission velocity, which is based on a coordinate system moving with the falling jet, depends on the concentration of the polymer solutions but not affected by the flow rate and the history of deformation in reservoir and nozzle. The transition point of deformation reported in our previous paper corresponds with the position where the upward relative transmission velocity coinsides with the velocity of the falling jet.

View full abstract

Linear stability of the spherical Couette flow is numerically investigated when the aspect ratio β=0.5. The present report gives a deteiled review of the numerical method to calculate the unperturbed flow and to solve the eigenvalue problem for the perturbation flow associated with the linear stability problem. Collocation method together with the Newton-Raphson method and the double QR method is adopted. In the course of convergence check, it is found that insufficient spatial resolution causes a serious underestimation of the marginal Reynolds number.

View full abstract

The flow through a curved rectangular duct with the curvature=0.1 is investigated numerically at the Dean number Dn=100 for various aspect ratio of the duct cross section by use of the spectral method of the polynomial function expansion. The flow is assumed to be uniform in the direction of the duct axis. Steady solutions which are symmetric with respect to the mid-horizontal plane of the duct cross section are obtained by the Newton-Raphson's method and their linear stability is studied. The bifurcation diagram of the solution shows that there exist three solution branches, the main vortex solution, the multi-vortex solution and the singular vortex solution. Some of the secondary flow patterns found by Akiyama et al.⁽¹⁾ are allotted to the solutions in the bifurcation diagram.

View full abstract

Passenger car torque converters have been designed with an increasingly lower profile in recent years for the purpose of achieving weight savings and a more compact size. However, flatter designs have tended to result in reduced hydrodynamic performance. The major cause of this performance deterioration is the pump, where relatively higher loss is generated at the core and suction surfaces. In this study, the internal flow characteristics of torque converters having pump impeller passages with four different contraction ratios were investigated numerically using a CFD code. The results show that there existed an optimum contraction ratio of the pump meridional passage in terms of efficiency. The four torque converters were then manufactured and tested in order to evaluate the change in their overall performance relative to the contraction ratio. The measured overall efficiency and visualized flow patterns tended to coincide well with the computation results.

View full abstract

A laser Doppler velocimeter (LDV) was used to investigate complex three-dimensional unsteady flow fields in a three-element automotive torque converter. The refractive-index matching method, which allows matching of the refractive index of the structural material to that of the working fluid, was successfully applied to flow measurement in cross sections in the pump and turbine impellers. Preliminary experiments were conducted to determine the temperature of the working fluid and the density of the tracer particles. Data were acquired synchronously with impeller rotation, making it possible to obtain the circumferential variation of the flow velocity. The results show that reverse flow in the circulatory velocity was observed in some cross sections in the pump. It was also shown that the secondary flow on the measured plane varied considerably from one cross section to another. The accumulation of low energy fluid by the secondary flow well explains the flow separation and reverse flow behavior.

View full abstract

Thi paper describes the experimental study of the radial supersonic flow behavior over the riblet surface mounted on the disk-shaped wall to reduce a pressure loss. The experiments have been performed in an indraft supersonic wind tunnel installed with disk-shaped channel at Mach numbers ranging from 1.8 to 3.3 just in front of a pressure rise. The static pressure rise in the pseudo-shock region observed in a channel with riblet surface became milder than that for the case of smooth surface, and significance of its difference was indicated by uncertainty analysis entimated at 95% coverage. The contours of stagnation pressure obtained by traversing a Pitot-tube showed that a higher stagnation pressure was maintained near the riblet surface compared with that for the smooth surface. Furthermore, it was found that the stagnation pressure loss reduced about 20% near the riblet surface at a Mach number of 2.5.

View full abstract

Orientation profiles of fibers in fiber suspension flow between two parallel plates with a square cylinder on the center axis was studied by using the computer image analysis. The orientational order parameter S decreases near the square cylinder because the flow changes its direction and also a wake is developed, while it is large near the channel wall owing to high shear deformation. The preferred angle and S decrease in the downstream vicinity of the square cylinder, and reach fully-developed values at a farther area. S along the channel center axis increases with increasing a fiber concentration and the Reynolds number. The size of a square cylinder has an effect on the fiber orientation near the cylinder. Since a stagnation region and a large wake are developed before and after a cylinder for a viscoelastic fluid, a lower value of S near the cylinder is obtained for the viscoelastic fluid compared to that for a Newtonian fluid. At a downstream area, however, the viscoelastic fluid gives a higher values of S.

View full abstract

This paper deals with an experimental study on the unsteady cavitation performance in a long cylindrical choke with the sharp-edged corner. The unsteady flow takes the trapezoid form with time at the downstream of a choke. Experiments are carried out for four kinds of the choke diameter d under various rates of the pressure changes while keeping the choke length l and the upstream pressure p₁ constant. The main objective of this study is to find four kinds of the critical cavitation numbers k_ci, k_cci, k_cd and k_ccd for three each location within a choke and only k_ci, k_cd for the outlet of choke : the incipient cavitation number k_ci, the desinent one k_cd, the cavitation number k_cci of "beginning of choking", which is the value at the moment which the full bubbles begin to appear in a choke, the cavitation number k_ccd of "end of choking", which is the value at the moment which the full bubbles in a choke begin to collapse. It has been clarified that each value of four kinds of these numbers depends on each location within the choke, the ratio of pressure change and the choke diameter.

View full abstract

A peristaltic flow is solved by finite difference method. The flow in peristaltic pump is three-dimensionally by the effect of the side walls. Since the actual peristaltic flow has a secondary flow in its cross section against the streamwise direction, so the three-dimensional flow analysis should be employed. In order to solve this problem, an algorithm for the solution of the incompressible Navier-Stokes equations in three-dimensional generalized curvilinear coordinate is presented. This method uses a third-order upwind scheme for the convective term and the second-order centered difference scheme for the viscous term. The Navier-Stokes equations and Poission equation of the pressure are solved by the plane Gauss-Seidel iterative method. The high-order numerical simulation of the peristaltic flow was carried out with 81×41×41 mesh points. As results of this investigation, three-dimensional flow structure of the peristaltic flow is cleared. The velocity in the plane perpendicular to the streamwise direction show the existence of the rotating vortices.

View full abstract

This study is concerned with the turbulent structure of three-dimensional boundary layer on a thin cylinder rotating in a uniform stream. A ratio of the turbulent shear-stress to the turbulent intensity, that is, a structure parameter, is significantly larger than the turbulent boundary layer on a stationary cylinder which has nearly the same value as in two-dimensional turbulent boundary layer. Terms appearing in the equations for the turbulent energy of each component of fluctuating velocities are estimated, and their roles in the energy budgets in this boundary layer are clarified. Particularly, the importance of redistribution terms and exchange terms between v-and w-energy is reconfirmed. Cross-spectra between u-and w-fluctuating velocities are examined across the boundary layer. The distribution of co-spectra and quad-spectra (i.e., the real and imaginary parts of the cross-spectra respectively) shows the existence of large-scale organized structure in this turbulent boundary layer.

View full abstract

Laminar free convection boundary layer flow along a porous wedge which is maintained at a uniform heat-flux have been studied theoretically. Boundary layer equations for the basic steady flow are solved numerically using an efficient finite-difference method in combination with an iterative method. The temporal neutral stability theory for wavelike disturbance of Tollmien-Schrichying type are then presented for the velocity and temperature functions. The corresponding eigenvalue problem for the disturbance amplitude functions is also solved numerically. The neutral stability curves and the disturbance amplitude of the velocity or temperature are given for some values of wedge angle parameter m, suction/injection parameter ξ and modified Grashof number G. The Prandtl number Pγ is taken to 0.73 throughout this paper.

View full abstract

Since velocity vectors obtained by Particle Imaging Velocimetry are distributed discretely depending on the number density of tracer particles seeded in the objective flow, a new estimation algorithm to reconstruct the whole-field flow has been proposed using the Laplace equation for velocity vector and the Poisson equation for velocity correction potential. Performance of the algorithm was examined with two-dimensional vortex flow, and the validity was evaluated quantitatively by two-dimensional correlation coefficient. The present algorithm was applied to torque converter internal flow, and the results reveal that the present method is applicable for various conditions to analyze the whole-field flow structure.

View full abstract

Turbulence structure in rising bubbly flow driven by bubble buoyancy was analyzed by using the Eulerian-Lagrangian equation model. A new calculation method for two-way interaction between dispersed bubble and liquid phase was proposed to simulate accurately the local fluctuation of liquid flow induced by bubble migration. The present two-dimensional simulation revealed that, (1) the turbulence structure had the longer wave length in vertical direction in case of the larger bubble, (2) the wave spectrum of kinetic energy of the liquid phase depended sensitively on the average void fraction and the lift force of bubble.

View full abstract

A low-Reynolds-number second-moment closure which completely eliminates wall-reflection redistribution terms is tested in turbulent flows with wall blowing and suction. The turbulence model was previously shown to give good predictions for fully-developed channel flow, boundary layers in adverse and favorable pressure gradients, and plane and round jets. In the present study, the model is used to calculate three boundary layers with blowing, two boundary layers with suction, and a channel flow with blowing and suction. The predictions are generally in good agreement with experimental data or direct numerical simulation data. The result encourages further testing of the model in various turbulent flows to establish a second-moment closure without wall-reflection redistribution terms.

View full abstract

Turbulence intensities in Couette·Poiseuille flow, developed between stationary and moving walls, have been measured by I-and X-type hot wires. The intensities in the wall region are affected by non-dimensional shear stress gradient parameter μ (≡u³_*/av), but not by Reynolds number Re^* (≡hu_*/v). As|μ|decreases, distributions of streamwise and wall-normal turbulence intensities shift upward or downward from those of plane-Couette flow depending on the sign of μ. In the turbulent core region, turbulence intensities of Poiseuille-type flow distribute quite differently from that of Couette-type flow. The effective parameter in this region is β, but the effect of β on the turbulence intensities is obscured by the low Reynolds number effect.

View full abstract

Turbulence statistics in Couette·Poiseuille flow are obtained by measurements. These include correlation coefficient, skewness and flatness factors and four-quadrant analysis of Reynolds shear stress-puv^^-. In the region of y⁺&le;30∼40, the distributions of all these quantities are only affected by non-dimensional parameter μ(≡μ³_*/av), as the mean velocity and the turbulence intensities profiles are. The four-quardrant analysis shows that the fractional contribution from 4th-quadrant is affected largely by parameter μ whereas that from 2nd-quadrant remains unaffected. In the case of 0<μ&le;94, the fractional contribution from 4th-quadrant is greater than that from 2nd-quadrant, unlike the conventional wall turbulent flow.

View full abstract

As a device in order to obtain a fully developed turbulent pipe flow as early as possible, a ringtype manipulator with a rectangular cross section was used. The experiment was performed in the situation where an air was flowing at Reynolds number of 9×10⁴. Ring-type manipulator were placed immediately downstream of a bellmouth entry at 3 positions, which are on the pipe wall, the outer edge of turbulent boundary layer and in the potential core region. The mean flow field behind the manipulator in a smooth pipe has been investigated by measuring the wall static pressure and longitudinal mean velocity. From the results, the streamwise change in the boundary layer characteristic values suggests that the recovery distance from the ring disturbance is shortest when the manipulator is at the outer edge of the turbulent boundary layer.

View full abstract

Acoustic resonant vibration occurs in cavity with tube bundle when flow velocity reaches at a certain velocity. Mode of the vibration is usually transverse to flow direction and way of prevention is not very difficult. However, in recent years, vibration of parallel mode sometimes occurs. Cause of the vibration must be an increasing of flow velocity and complication of duct shape. There is no effective method to prevent the vibration so that it is very important to find out a method of judgment in design stage, whether the acoustic resonant vibration will occus or not. Vortex shedding is influenced by the acoustic vibration and the vibration is also influenced by the vortex shedding, so the acoustic vibration seems to have feed back loop. Mechanism of the vibration having a feed back circuit is discussed in thip paper.

View full abstract

Separated flow in a centrifugal impeller under the unsteady operating condition was examined by a discrete vortex method. Calculations were carried out in the case of impulsive change of the rotational speed at the instant τ=0 (τ : non-dimensional time) under a constant flow rate. Results shows that the total head rise of the impeller obtained from the unsteady Bernoulli equation responds quickly, and reaches to the final value before τ=1.

View full abstract

There are two typical cases in the problem of the two-dimensional unsteady flow past a rotating circular cylinder which is impulsively started. (1) The circular cylinder is impulsively started with rotation from rest and (2) the rotating cylinder is impulsively started. The difference between these two cases is given by the far-field condition. A vortex method is powerful for these problems because the far-field condition is naturally imposed, so the hybrid vortex method combined with a panel method, whose numerical accuracy was tested on low Reynolds number flow around an impulsively started cylinder without rotation in the previous work, is used in the present paper. The detailed results of the flow around the rotating cylinder with a constant anticlockwise angular velocity are shown and we see that there is difference of flow between these initial conditions and its difference does not become clearly small with development of time in the case of high angular velocity, in particular, the difference in drag force does not become clearly small with time although it in lift force becomes small with time.

View full abstract

The characteristics of the near-wall grid dependency of the low-Reynolds-number linear and nonlinear eddy viscosity turbulence models have been examined in fully developed wall shear flows. The tested ten models consist of recently proposed k-ε, k-ω, q-ζ and k-ε-A₂ models. Although the recent models include complicated model functions to mimic the near-wall behaviours of turbulence quantities, they showed less sensitivity to the grid-spacing than the traditional near-wall model. These models perform very well on the grid with y⁺&ap;2 of the first grid-node distance from the wall, whilst y⁺&ap;4 is sufficient for some models. From the analysis of the discretization error, it has been explained that the selection of the transport variables behaving linearly near walls does not directly contribute to decreasing the grid sensitivity, but in some cases, the grid sensitivity is affected by the accuracy of the predicted behaviours of turbulence quantities. It has also been pointed out that the models with the w transport equation require adequate grid spacing near the wall which is neither too fine nor too coarse.

View full abstract

As the use of wind turbine generator system (WTGS) expand over the world, complex terrain has become the site-candidates more often. Therefore, "Complex Terrain" is now one of the keen issues, which however, has been the original problem in a mountainous country such as Japan. International Electrotechnical Commission (IEC) is now actively preparing a standard entitled "Power Performance Measurement", which describes the procedure of measuring performance of a wind turbine generator system. However, the standard is applicable to WTGS built on rather flat open areas in principle and is hardly applicable to WTGS in complex terrain. This has been well recognized, but it is the opinion of the majority of the IEC working group that understandings and technique for power performance measurement of a WTGS in complex terrain is still under development. Therefore, a new and more reliable method of performance measurement of a WTGS in complex terrain is required not only from the technical aspects but also from economics. The particular purpose of the present work is to measure and evaluate the wind speed and power performance of the national 500 kW research WTGS, while the generalized purposes are : (1) to develop a reliable method of wind speed measurement and (2) to develop a reliable method of power performance measurement of a WTGS in complex terrain. In this work, firstly, the measured wind data in typical complex terrain such as Tappi wind-park are analyzed. Secondly, a new method of performance measurement/analysis is discussed. Method of bin is applied for data analysis, while a reference meteorological mast and reference factors are newly introduced. Finally, the new method is evaluated.

View full abstract

This paper describes about thermal design and simulation method of closed cabinet applying heat exchanger for inner air cooling. In controller system which we used for this research, the cabinet is equipped in outdoor, and the cabinet is sealed to prevent the inducement of dust. This system dissipates 630 Watts in total, so it had a problem that the inner air temperature of the cabinet was raised over 30 K. So we equipped a heat pipe heat exchanger in the cabinet, and we also changed the air flow path to realize the best performance of heat exchanger according to the results from thermal and fluid simulation using macroscopic model of heat exchanger to reduce computational load. In conclusion, inner air temperature rise had reduced below 15K, and the result was in good agreement with the simulation.

View full abstract

Numerical simulations have been made on in-line elliptic tube banks at low Reynolds numbers. Details of the flow and heat transfer characteristics in the tube banks are clarified for the fully developed flow state under the condition of uniform heat flux. The pressure loss coefficient, the local and mean heat transfer coefficients are presented along with the flow and temperature fields. The effects of the longitudinal and transversal pitch ratios and the axis ratio upon them are made claer. It is found that an increase of the longitudinal pitch results in an increase of the pressure loss and the heat transfer. On the other hand, an increase of the transversal pitch causes a decrease of them, especially the pressure loss. The axis ratio greatly effects these characteristics. These results suggest that an appropriate arrangement of the elliptic tube may enhance the heat transfer performance of the tube bank.

View full abstract

Heat transfer and CHF (Critical heat flux) characteristics of flow boiling of R-113 in helically coiled tubes were experimentally investigated. Two coiled tubes with coil diameters of 0.165 and 0.32m, and 10mm I.D. were tested at a pressure of 0.39 MPa. In the nucleate boiling region, circumferential difference in heat transfer was not clarified qualitatively as well as quantitatively. The ratio of circumferential average boiling heat transfer coefficient to that of the single-phase flow in a curved tube, h_TP/h_LO, was found to be rather close to the correlations of Pujol-Stenning and Ueda-Kim for a straight tube flow. In the high quality region, the heat transfer coefficient was highest at the coil outside and lowest at the inside, and the average heat transfer coefficient ratio h_TP/h_Lo was expressed in terms of Martinelli parameter, X_tt, only. A correlation applicable to both the nucleate boiling and forced convective evaporation regions was proposed. The CHF of the coiled tube was higher than that of the straight tube when the coil diameter was small, but it became considerably low at low mass velocity when the coil diameter was large.

View full abstract

Impingement heat transfer and flow along the radial and circumferential direction by a single circular laminar jet with a confined insulated wall were estimated numerically in three-dimensional form and were recognized by visualization of a thermosensitive liquid crystal. Local heat transfer is divided into three regions. The first is two dimensional forced convection region whose structure is coaxial circle. The second laminar mixed-convection region begins with onset of the buoyancy-driven flow, which corresponds to thermal plumes rising from the heated impingement wall at discrete circumferential locations. The ascending and descending pair fluid forms respectively longitudinal streak lines in the mixed convection region. The number of pairs is found to depend on the aspect ratio of circumferential length and distance between nozzle and impingement surface. The local value of Gr/Re² on the onset of the buoyancy driven flow is from 80 to 100. The third is like Benard convection which constructs several cells. The radial Nusselt number distribution averaged along the circumferential direction is also presented corresponding to flow.

View full abstract

The evaporation patterns of methanol liquid film feeded on heated rotating tube inner wall is visualized by optical image de-rotation method, while the heat flux is measured under the same condition. It is shown that there are several flow patterns of natural convection and that the evaporation heat transfer coefficient of the natural convection film increase with the rotational acceleration but the nuclear boiling is restrained within the conditions of 15 to 325 for the ratio of rotational acceleration and gravitational one, and 0.8 to 4.2 mm for film thickness, 2500 to 75000 W/m² for heat flux, 5 to 40 K for wall super heat.

View full abstract

The pseudo-dropwise condensation driven by concentration difference, which is called "Marangoni Dropwise Condensation" in this paper, appears in the "positive system" of the condensation of vapor mixture. The observation and measurement were performed for Marangoni dropwise condensation of water and ethanol vapor mixture. The processes of initial droplets formation from thin condensate film and drop departure were taken by using a high speed digital camera simultaneously with the heat transfer measurement. The variations of the initial drop distances exhibited V-or U-shape against the surface subcooling. Their minimum values were in the range of 0.03mm∼0.15mm depending upon the etanol concentrations. The surface subcooling in which the minimum initial drop distances appeared coincided with those for maximum heat transfer coefficient. Also the variation of departing drop diameters had similar trend. Since the augmentation of driving force of Marangoni dropwise condensation made the initial drop distance and the departing drop diameter smaller and the condensate film among droplets thinner, the heat transfer was enhanced.

View full abstract

In order to clear the mechanism of the vapor explosion, two kinds of experiments were performed. In the experiment 1, an upper side-open transparent vessel filled with molten metal (solder, tin, lead and zine) was immersed in the water tank. It was found that vapor cavities were generated in the metal and the vapor bubble connected with a cavity collapsed just before the occurrence of a vapor explosion. In the experiment 2, the apparatus consists of a water tank and a heated vessel with a cavity. Subcooled water in the water tank flows into the vapor cavity, immediately after the pressure in cavity is abruptly reduced because of the vapor condensation. Then the pressure rises rapidly by the vaporization of water in the cavity. A chain of phenomena, the generation of vapor cavities, the pressure descent due to vapor collapse, the water inflow into the cavity, and the rapid pressure rise, can be a part of the process of vapor explosion. The relation between the metal temperature just before the occurrence of a vapor burst and the minimum film boiling temperature criterion of Dhir & Purohit was investigated.

View full abstract

In order to improve a boiler efficiency, latent heat recovery from the flue gas is very important concept. Condensation heat transfer on horizontal stainless steel tubes was investigated experimentally by using an actual flue gas from a natural gas boiler. The experiment was conducted at different air ratios of the flue gas and wide range of tube wall temperatures. The condensation pattern was similar to a dropwise condensation near the dew point. Decreasing the wall temperature, the wall region covered with a thin liquid film increased. The heat and mass transfer behavior was well predicted with the analogy correlation at the high wall temperature region. At the low wall temperature region, the total heat transfer was higher than that predicted by the analogy correlation.

View full abstract

Natural convection heat transfer between concentric rectangular parallelepipeds was studied numerically for low Rayleigh numbers Ra (≤3500) with aspect ratios of the inner parallelepiped of 2.0, 4.0, 6.0 and 8.0. It has been found that the flow patterns for the higher Rayleigh numbers in the space over the inner parallelepiped are ring or rectangular rolls. The number of rolls increases with the aspect ratio. The flow pattern in the side space is an oblong circulation, which extends into the bottom space. The local Nusselt number distribution on the top surface of the inner parallelepiped has peaks at the stagnation points. The relation between the Nusselt and Rayleigh numbers on the top surface is similar to that of the Rayleigh-Benard convection obtained by Silveston [Chandrasekhar. S., Hydrodynamic and Hydromagnetic Stability, (1961), 68, Oxford University Press], while on the side and bottom surfaces the Nusselt number increases proportionately with the power of the Rayleigh number.

View full abstract

Numerical simulations of heat transfer in a furnace of practical scale were carried out, coupled with a prediction of the flame. Radiative heat transfer was computed taking its nongray characteristic into account. Convective heat transfer to the furnace wall was incorporated as well. The numerical computations were conducted based upon the conservation equations of mass, momentum, energy and species together with k-ε turbulence model and eddy dissipation model for estimating combustion rate. Radiative heat transfer was computed with the radiant heat ray method in combination with a weighted-sum-of-gray gas model or a wide-band model. Predicted results were compared with the experimental data. Comparisons showed that the profiles of temperature and species concentration can be predicted fairly well and that the weighted-sum-of gray gas model gives a good prediction of heat flux to the furnace wall as well as the wide-band model. Also the influence of soot on the furnace heat transfer was estimated.

View full abstract

One of the restrictions in the thermal design of the boiling water reactor (BWR), is pointed out to be the lack of heat removal from the nuclear fuel rods near the supporting spacer. For this the spacer must be carefully designed focusing attention on the geometry of the spacer. The purpose of the present paper is to experimentally investigate the effect of the spacer geometry on the film breakdown near the spacer. Both ring and grid type spacers were used in the present experiment. From the frequency of the dry patch formation and the temperature distribution near the spacer, it is clarified that the ring type spacer is superior to the grid type spacer.

View full abstract

The dissipated effective kinetic energy in the refrigeration cycle can be recovered using the two-phase ejector. Performance of the two-phase nozzle plays an important role in developing the two-phase ejector for the refrigeration cycle. It is also important to know the efficiency of the two-phase nozzles for refrigerant because the physical properties of refrigerant are different from that of hot water which has been studied. The purpose of the present study is to elucidate the nozzle performance of two-phase flow of refrigerant R134a which is composed of no chloric radical. The experimental results shows that the efficiency of the long nozzle of about 110 mm is high up to 90% which is higher than that using hot water. The efficiency of the nozzle which has the same outlet diameter as an ideal nozzle is higher than others, because the flow flush out of it is at the most suitable expansion condition.

View full abstract

A molecular dynamics simulation has been performed for the two-dimensional Lennard-Jones (12-6) fluid. Saturated liquid at low temperature, compressed liquid, and fluid at supercritical temperatures with various densities were analyzed. Number of bonds per molecule, lifetime of bonds, connectivity of molecules to form clusters, stability of clusters, and their responces to temperature and density are compared with those of hydrogen bond in water.

View full abstract

This paper presents double-diffusive convection in a square cavity filled with binary gas, due to horizontal opposing temperature and concentration gradients. The effect of Lewis number was considered under the conditions of Prandtl number Pγ=1, buoyancy ratio N=1 and thermal Rayleigh number Ra_r=10⁴ and 10⁵. Numerical solutions are obtained by a Chebyshev collocation technique with high resolution. Depending upon the Lewis number, three kinds of flow structures are identified, i.e., symmetric steady flow, asymmetric oscillatory flow and symmetric oscillatory flow. Oscillatory flow occurs in the regime of thermal dominant flow and it leads to a periodic change between stable and unstable states in species stratification due to the thermo-solutal interaction.

View full abstract

The coherent structure and relevant heat transport in geostrophic flows under various density stratification has been studied by using both direct numerical simulation and rapid distortion theory. It is found that in a neutrally stratified flow under system rotation, the temperature fluctuations become very close to two-dimensional and their variation is very small in the direction parallel to the axis of rotation. Under the stable stratification, the velocity and temperature fluctuations tend to oscillate with the Brunt-Vaisala frequency. Under the unstable stratification, on the other hand, vortex columns are formed in the direction parallel to the axis of rotation. However, the generation of the elongated vortex columns cannot be predicted by the rapid distortion theory. The non-linear term is required to generate these characteristic vortex columns.

View full abstract

The purpose was to elucidate the production process of cerium dioxide particles, and the influenc of the production conditions on the pyrolysis process of cerium nitrate droplet and the powder characteristics of the produced particles were examined. Particularly, initial droplet diameter, initial metal concentration and air temperature were attracted as the production conditions and the influence of them on the size distribution, specific surface area, bulk density and shape of the produced particles were investigated. The mono-dispersed spray with narrow droplet size distribution was injected into a hot air stream and cerium dioxide particles were collected. The size distribution of produced particles was wide in comparison with the initial droplet size distribution, because the sample droplets were scattered or broken up by foaming in the process. The scattering or breakup of droplets due to foaming was promoted with increase of air temperature, the frequency of fine particles increased and mode diameter of produced particles dicreased. From the observation with SEM and TEM, the produced particles were porous.

View full abstract

Direct Internal Reforming Molten Carbonate Fuel Cells (DIR-MCFC) have such several problems as cross gas-leak between the channels and inhomogeneity of the gas flow distribution since the DIR-MCFC pack a catalyst into the anode channel. In order to solve those problems, it is necessary to elucidate behaviors of the pressure drop in the DIR-MCFC and develop the simple prediction method. As a result, two equations of friction coefficient in the DIR-MCFC under the conditions with and without the catalyst were experimentally obtained, based on Wieting's equation. The simple prediction method was developed by means of comination of the two equations.

View full abstract

An experimental study was carried out to make a quantitative analysis of fuel droplet diameter at a high pressure and temperature occurring in combustion systems. A fueling system was developed to atomize fuel into droplets having a single small diameter. By using this system, fueling was made into a visualized device where high pressure and temperature nitrogen flew under steady-state conditions. The laser induced fluorescence method was utilized to determine the n-dodecane fuel and iso-octane fuel droplet behavior and the phase doppler particle analyzer to measure the diameter and velocity of the droplets. As a result, quantitative analysis of a fuel droplet got possible. The results indicate that due to its lower boiling point, iso-octane droplets tend to evaporate much faster then dodecane droplerts.

View full abstract

In order to examine the ignition process, direct numerical simulation was carried out on spontaneous ignition of fuel gas jets injected parallel to a steady hot air stream. Full kinetics was used in the calculation of chemical reaction. Hydrogen was chosen as the fuel gas because of its relatively simple chemical reaction mechanism. Calculated results revealed the follows. (1) The ignition point moves from the peripheral region in the upstream of a fuel jet to its top region with a decrease in spouting velocity of fuel jet, and an increase in air temperature. (2) The ignition takes place in fuel-lean region situated at the end of hot air side in the mixing layer, where turbulence is weak. (3) The assumption that Lewis number is unity is not suitable, because the effect of molecular diffusion is significant in the ignition of fuel gas jets.

View full abstract

Fuel jet diffusion flame flow was simulated by parallel computing on two workstasions of NEC 4800 series with CPU chip of R 4400. MPI, Massage Passing Interface was employed for communication between them. The domain decomposition technique plays an important role in increasing the effciency of parallel processing. In this paper, the dynamic subdomain boundary moving method, where the subdomain boundary is moved to optimize the load balance of each workstation, was applied to calculate a combustion flow. As a result, it was found that the present method was effective when the load balance was disturbed by other users' processes.

View full abstract

Experimental and numerical studies are made of transient H₂/N₂-air counterflow diffusion flames unsteadily strained by an impinging micro jet. Two-dimensional temperature measurements by the laser Rayleigh scattering method and numerical computations taking into account detailed chemical kinetics reduce the following results. (1) Transient local flame extinction is observed where the micro jet impinges. The transient flame can survive instantaneously in spite of the quite high stretch rate where the steady flame cannot exist. (2) Reignition is observed after the local extinction induced by an impinging micro air jet. The predicted behaviors of the local transient extinction and reignition are well confirmed by the experiment. (3) The reignition is hardly observed after the extinction by the micro fuel jet impingement. (4) These characteristic phenomena of (2) or (3) are dominated by the concentration or dilution of H₂ due to the preferential diffusion with the concave or convex flame curvature, respectively. (5) The maximum flame temperature cannot be rationalized by the stretch rate but changes widely depending on the unsteadiness and the flame curvature in relation to preferential diffusion.

View full abstract

The cross-sectional images of turbulent premixed flames of homogeneous fuel-air mixture in an engine cylinder were obtained by a laser tomography, and the fractal characteristics were investigated. It was observed that in an engine cylinder under high pressure condition, the turbulent flame was a smaller and more complicated structure, comparing with that under the lower pressure condition in a closed combustion chamber. The fractal dimension increases with the increase of the turbulence intensity and mixture density. The fractal dimension is expressed as a function of the increase ratio of mixture density and the ratio of turbulence intensity to laminar burning velocity. The inner cutoff scale of turbulent flames is expressed as a function of Karlovitz number.

View full abstract

A stability limit of the diffusion flame with fuel injection from a porous wall in a laminar flat plate boundary layer is measured as functions of fuel (CH₄) concentration of CH₄/N₂ injectant mixture (x) and its injection velocity (v). The free stream velocity (U_∞) is set as 0.6 m/s. The thermal condition at the wall is controlled by setting temperature at the upstream end of the porous wall as a reference temperature. When v>20 mm/s, the flame becomes unstable with the separation of leading flame edge with decreasing x. The value of x at the stability limit is constant without regard to v as long as the wall temperatrure is kept constant. As the wall temperature is decreased the value of x increases. The separation is supposed to take place as a result of the limit of the reaction rate. When v<20 mm/s, the flame becomes unstable with the oscillation. The value of x at the stability limit increases drastically with decreasing v. The oscillation takes place mainly due to the repeat of the extinction due to heat loss to the wall and the flame propagation in the combustible layer.

View full abstract

The influence of hydrodynamic instability on the structure of cellular flames is investigated using numerical method. The equation used is the compressible Navier-Stokes equation including an one-step chemical reaction. We superimpose sufficiently small disturbances on the plane flame and calculate the evolution of flame fronts to obtain the relation between the growth rate and the wave number, i. e., the dispersion relation. As the flame temperature increases, the growth rate becomes larger. The unstable range hardly changes, and the peculiar wave number is almost constant. Thus, the spacing between cells in the cellular flame is independent of the flame temperature. Moreover, we superimpose a considerably large disturbance with the peculiar wavelength to study the structure of cellular flames. The higher the flame temperature, the deeper the cell and the broader the flame surface. In addition, the cell depth and the surface area of the three-dimensional flame are larger than those of the two-dimensional flame.

View full abstract