The cavitation erosion problem in liquid flow systems is investigated by many engineers and scientists from the viewpoints of engineering applications as well as basic scientific interest. At the present stage, the mechanism of cavitation erosion is not completely elucidated. One of the most important points is that most investigations have not been conducted under real flow conditions. In the present investigation, simultaneous measurements of cavitation bubbles with impulsive force were realized in flow using a new system combining a high-speed video camera with an impulsive pressure sensor. Thus, some new findings have been revealed concerning the characteristics of bubble behavior near the bubble collapsing period.

View full abstract

This paper describes an experimental study on the process involved in the disappearance of arch vortices shed from a square cylinder of finite length, and the accompanying changes in the turbulent near wake, as a function of the aspect ratio of a square cylinder. The experiment was carried out in a circuit-type wind tunnel having a 200 mm×200 mm working section of 2000 mm length at the Reynolds number Re of 9200. The time-mean velocity and turbulence intensities were measured by a laser Doppler velocimeter. Vortex formation was observed by flow visualization in a water channel at Re=4600. Consequently, details of the flow pattern in the turbulent near wake behind a square cylinder were obtained, such as information on shedding arch vortices and horseshoe vortices. The aspect ratio for generating arch vortices seems to be higher than H/D=0.7, whereas the horseshoe vortices are generated around a square cylinder for all aspect ratios studied here.

View full abstract

Experimental studies on drag reduction of a flat plat normal to airstream were carried out. To control the flow around the plate, a small rod was set upstream of the plate. The chord length of the plate, D, was 50 mm and the diameter of the rod, d, ranged from 2 to 20 mm, and the distance between the axes of the rod and plate, L/D, was varied from 1.0 to 2.6. The Reynolds number ranged from 1.9×10⁴ to 7.7×10⁴. For the case without vortex shedding from the rod at d/D=0.4 and L/D=2.0, the maximum reduction of the total drag coefficient is about 20 to 30% compared with the drag without the rod in the same range of the Reynolds number.

View full abstract

Mean velocity profiles of Couette·Poiseuille-type turbulent flow developed between fixed and moving walls have been measured. The relevant parameters for the wall, 1/2-power and core region are μ(=μ³_*/αν, where α=1/ρ·dτ/dy) and Re^*'(=u_*h'/ν), β'(=αh'/u²_*) and Re^*', and β(=αh/u²_*) and Re^* (=u_*h/ν) respectively. The effects of these parameters on the velocity profiles in each region considered. Except for the core region of the Couette-type flow, each region has a reference wall. In the wall region, the wall law varies largely with μ but slightly with Re^*'. Typically, the additive constant B of the logarithmic-velocity low (or Van-Driest damping factor A⁺) is shown to depend only on μ. The 1/2-power velocity law depends not only on β', as well known, but also Re^*' under low Re conditions. In the core region, the normalized eddy viscosity ε_T/u_*h' is constant irrespective of Re^* and β for Poiseuille-type flow while it depends on both Re^* and β for Couette-type flow. This indicates that the turbulent structures of the core region are quite different between Poiseuille-and Couette-type flows.

View full abstract

In the previous paper, we showed that the flow rate distribution into the branch tubes from the inlet header of a 2-tube pass multitube heat exchanger which has the tube bundle arranged in a rectangular shape is markedly different from that which has the tube bundle arranged in a square shape. In the present experiment, to clarify the mechanism which causes this difference, the flow pattern in the inlet, outlet and return header of the 2-tube pass multitube heat exchanger was visualized using aluminum powder as a tracer, and the pressure distribution on the bottom wall of the inlet header was measured. The results show that the flow rate distribution is closely related to the flow pattern in the inlet header. Also, the eddy generated in the return header is strong and contains air bubbles which are never discharged.

View full abstract

Recent computer development has enabled such a numerical simulation of turbulence as the direct numerical simulation (DNS). Direct numerical simulation provides valuable information of the flow field, such as the distributions of fluctuating velocity, pressure, and vorticity in the three dimensional space, which are difficult or often impossible to be obtained from experiments. In this study, the unsteady behavior of coherent structures in the two dimensional turbulent channel flow reproduced by a direct numerical simulation was visualized and animated with the aid of three dimensional computer graphics technique. The space-time relationship between vortical motions and streaks, ejections, sweeps was clearly observed from a moving frame at a speed of U⁺=14.0. In addition, it is found from the spectrum analysis that the coherent structures are constructed with relatively low wave number components of the fluctuating velocities and pressure.

View full abstract

Thermohydraulics have played an important role in a variety of industrial devices. In particular, constitutive equations have been developed to describe the interactions of gas and liquid at the interfaces of various flow patterns. Such behavior used to be solved numerically by a set of finite difference equations. However, numerical simulation of interfacial motion in a flow is still a incomplete in many cases. The cellular automaton method is expected to be an effective technique for simulating fluid motion numerically. Because of discretized space, time and variables, the discrete dynamic system can be defined, and the simple structure of dynamics allows it to be dealt with easily with a large scale of freedom. These enable us to realize discrete molecular dynamics on the computers. We have demonstrated that an automaton solution of two-component counter-flow results in turbulent behavior which scatters droplets at its interface. After evaluation of the turbulence intensity, velocity distribution of the averaged solution and an inertial energy, it can be concluded that this method is useful to simulate two-component flow.

View full abstract

In this study the plane wall jet under the self-preserving pressure gradient was measured to investigate the effect of the ratio of the maximum velocity U_m, to the external stream velocity U_e on the turbulent quantities. Velocity ratio U_m/U_e was set at three values within the range of 2.0 to 4.0. Under such conditions, three components of turbulent intensity and Reynolds shear stress were measured by using two linearized hot wire anemometers. The turbulent flow is closely self-preserving in the range of 57<x/S₁<162. However, the turbulent quantity profiles for each case may differ depending on the velocity ratio. The absolute value of each turbulent quantity in the inner layer decreases with decreasing velocity ratio ; conversely, in the outer layer, it increases. Several characteristic scales nondimensionalized with an excess velocity U₀ for both inner and outer layers have the same tendency independent of the velocity ratio.

View full abstract

The paper considers the modeling of the asymptotic near-wall behavior of the dissipation term in the Reynolds stress transport equation. A new form of dissipation model is proposed. In contrast to conventional near-wall modeling, the present model avoids the use of the wall-normal vector for generality. This model, unlike existing models, reproduces the exact gradient of the dissipation components ε₁₁ and ε₃₃ at a wall, where the indices 1 and 3 denote the streamwise and spanwise directions, respectively. The model and two representative existing models are evaluated in turbulent plane Poiseuille and Couette-Poiseuille flows with the aid of direct numerical simulation data. It is shown that the present model gives significantly improved dissipation behavior.

View full abstract

A Eulerian/Lagrangian-type numerical simulation was performed for two-dimensional fluidized beds, in which the particle motion is restricted by parallel front and rear walls. Particle motion was calculated using ordinary Newton's equations of motion, modeling the contact forces by the DEM (discrete element method). The local averaged equations were solved to calculate the fluid motion, taking into account the interaction between fluid and particles. Two kinds of calculations were performed regarding the calculation of particle motion : two-dimensional(2D) model and three-dimensional(3D) model. Results for both cases were compared and the effect of the walls on particle motion was discussed. Cases where partition walls are set up in the bed were also calculated.

View full abstract

A vectorial expression of the Leslie-Ericksen theory is transformed to a tensorial expression. The insensitivity of the nematic liquid crystals to a change of sign of a director can be reflected into the governing equations by the tensorial expression. As its application, drag flows of tumbling nematic liquid crystals between parallel plates are considered. For a homeotropic boundary condition, a phase transition with hysteresis is found for the maximum angle of the director in relation with the director tumbling. The change in director alignment largely affects the velocity profile. Furthermore, phase transition in the director alignment gives rise to a discontinuous jump in apparent viscosity.

View full abstract

The objective of this study is to examine mumerically the relation between the flow structure and the dynamic characteristics for a laminar separated flow through a two-dimensional symmetric sudden-expansion channel of expansion ratio α. A stream function-vorticity formulation was used and solved by the finite difference method using a pseudo-unsteady technique. For the cases of four different expansion ratios α=2, 3, 3.5 and 4, calculations were performed for the range of the Reynolds number 20≤Re≤400. The results showed that the critical Reynolds number Re_C (at which the asymmetric flow began) decreased with increasing α. Stream lines, velocity profile, distributions of pressure and pressure gradient were presented and discussed to clarify the relation between flow pattern and pressure gradient.

View full abstract

Two-dimensional fiber suspension flow between parallel plates was simulated using the Dinh-Armstrong constitutive equation for semi-concentrated fiber suspensions. Sixth-order and fourth-order closure approximations were also evaluated. The velocity profile for the fiber suspension flow is flat compared with Newtonian flow, because the preferred angle for fibers is large in the channel center region, resulting in an increase in the local viscosity. The order parameter decreases and the preferred angle becomes large with increasing fiber concentration. The velocity field is accurately predicted by the equations with quadratic closure approximations used in this study, whereas the approximations do not provide satisfactory results for the orientation field. With respect to the order parameter, sixth-order closure approximation gives slightly less accurate predictions than fourth-order approximation.

View full abstract

Flow regime, void fraction, rising velocity of slug bubbles and frictional pressure loss were measured for air-water flows in capillary tubes with inner diameters in the range from 1 to 4 mm. Although some flow regimes peculiar to capillary tubes were observed in addition to commonly observed ones, overall trends of the boundaries between flow regimes were predicted well by the Mishima-Ishii model. The void fraction was predicted well by the drift flux equation used together with a new equation for the distribution parameter as a function of inner diameter. The rising velocity of the slug bubbles was also predicted well by the drift flux equation. The frictional pressure loss was reproduced well by Chisholm's equation used together with a new equation for Chisholm's parameter C as a function of inner diameter.

View full abstract

The flow resistance of dampers in a duct is given in charts and tables in several databooks published by academic societies. It has often been indicated that some data are not reliable. This fact is inconvenient from the standpoint of use of the data for design. Hence, the flow resistances of parallel dampers in a duct of retangular cross section are discussed. Consequently, effects of the angle θ and the number N of dampers on the resistance coefficients are clarified quantitatiely. The resistance coefficients of dampers are given by the following general equations : ζ=1.93 N^-0.58[(1-β)/β²] for θ=10∼45°and ζ=7.3 N^-0.88[(1-β)/β²]^0.5 for θ=50∼70°, respectively, where β is the open ratio.

View full abstract

The saltation velocity (or the critical velocity) for coarse particles is predicted numerically, where the particle motion is traced by the Lagrangian approach by considering the fluid drag, gravity force, Magnus force, and Saffman force acting on the particles, whereas the gas flow is calculated by the marching method. The interaction between gas and particle is considered on the level of mean gas flow. It is assumed that the total pressure drop in a region of terminal velocity of particles consists of two parts, one due to the gas alone and the other due to the fluid drag. The fluid drag consists of the particle source term in the equation of motion for the gas, where the change in fluid drag due to particle-particle collision as well as particle-wall collision is considered. The result is quantitatively compared with several well-known empirical formulae for the saltation behavior.

View full abstract

A sputtering apparatus with a multiholed target, which has a large number of small holes on its surface, is proposed for the improvement of step coverage in a contact hole. The step coverage depends on the depth h of the hole drilled in the target and the deposition probability P_C on the side wall of the hole. The bottom wall coverage of the contact hole is improved considerably by the use of the multiholed target. The optimum aspect ratio is h/d=1.0∼1.5, where d is the diameter of the hole.

View full abstract

This paper describes the three-dimensional shock wave reflection over a corner of two intersecting wedges. A numerical study was carried out to investigate shock wave reflection in three-dimensional flow fields. The Weighted Average Flux (WAF) method was applied to solve three-dimensional unsteady compressible Euler equations. The results were compared with a numerical simulation using the Harten-Yee TVD finite difference scheme. The WAF method gave high resolution and was found to be useful in simulating this three-dimensional shock wave reflection without use of fine computational grids. The three-dimensional shock wave reflection was also analyzed with the shock polar method valid for two-dimensional obligue shock waves. The results obtained from the shock polar analysis agreed with the numerical simulations. A shock tube experiment was also conducted using double-exposure holographic interferometry. In order to obtain three-dimensional images of shock waves, the time interval between the first and the second exposures was set to be two or three micro seconds. This enabled clear observation of the structures of the three-dimensional shock wave reflection. Good agreement between this flow visualization and the results of the numerical simulations was observed.

View full abstract

A laser-induced fluorescence technique to diagnose a three-dimensional transonic flow with complicated shock wave-boundary layer interaction is presented. The diagnostic system consists of an argon-ion laser sheet traversing laterally in the flow field with seeded iodine as fluorescence material, a CCD camera with an image intensifier and a microcomputer for image processing. The temperature distributions in a rectangular duct with a swept-back bump are investigated by this system, and the structure of the flow field is clarified including the curved shock waves and the boundary-layer separation.

View full abstract

A numerical study of underexpanded axisymmetric supersonic jets is presented. The TVD finite-difference scheme is employed for the numerical calculation of Euler's equations. The numerical results are compared both with results from the theoretical shock cell spacing formula suggested by Tam and with previous experimental results. For nozzles having no divergence angle at the nozzle exit, a good agreement is obtained among the theoretical, experimental and numerical results for pressure ratios across the nozzle until the occurrence of a Mach disk in the jet. For nozzles having a divergence angle at the nozzle exit, the effects of the divergence angle, the nozzle design Mach number and the pressure ratio across the nozzle on the shock cell length are discussed and clarified.

View full abstract

Direct numerical simulations of spatially developing mixing layers were conducted to clarify the effect of phase difference of inflow disturbances on the growth of mixing layers. The disturbances consist of the fundamental and subharmonic modes of the most unstable wave based on linear stability analysis, and the energy transfer between mean flow, fundamental and subharmonic modes was examined. The results show that it is possible to control the development of mixing layers by controling phase difference between two modes. The phase difference determines energy transfer in the regions downstream of the roll-up location where energy is transferred from fundamental to mean flow, and upstream of the pairing location where the energy is transferred from mean flow to subharmonic. This mechanism causes the enhancement and suppression of the growth rate of mixing layers. The phase difference which produces the slowest development of a mixing layer is about 11/16π.

View full abstract

This paper presents experimental and numerical analyses for the flow characteristics of double coaxial pipe jets, which consist of a central round main air jet and a coaxial annular sub air jet, impinging normally onto a flat plate. As the results, it was found that the mean flow characteristics, namely each velocity profile, decay of maximum velocity, spread of jet, and pressure distribution, and the turbulence characteristics, namely each turbulence component and Reynolds shear stress, of a round impinging jet can be controlled by an annular coaxial sub jet. In paticular, turbulent kinetic energy near the flat plate is augmented considerably by the existence of a turbulent shear layer between main and sub jets. The increase in a forced convective heat transfer on the plate is also expected. Numerical analysis was carried out using the k-ε turbulent model.

View full abstract

A method for designing an optimal valve control is presented to suppress the shock due to the surge pressure caused by the rejoining of a separated column produced at various periods of an operationg cycle of a hydraulic machine press without reducing the cycle time. In this method, the time history of the input voltage signal for the valve, which controls a certain operation period (e. g., decompression period) in an operating cycle, is assumed to be of N+1 stage break-line mode from a specified initial value to a final one ; the mode being characterized by the coordinates of N break points. These points are determined by a computer program which incorporates a simulation routine for the system dynamics together with a direct search algorithm so that a minimum value of valve operation time can be achieved under the prescribed conditions of no column separation or the magnitude of surge pressure being lower than an allowable value. Experiments carried out on a small-sized model hydraulic test circuit dynamically similar to a real machine in the 20-MN class suggest that the present method is very useful for the development of a high-performance hydraulic machine press with low shock as well as with rapid operating cycle.

View full abstract

Noise generated by an axial fan is mainly composed of two types. One is rotational noise which has discrete frequency components, and the other is turbulent noise which has a broad band frequency component. Generally the pressure change due to blade passage causes rotational noise, and the pressure and velocity fluctuations cause turbulent noise. In order to reduce rotational noise, the discrete frequency components should be dispersed into many frequency components by adopting an unequally spaced configuration. For the reduction of turbulent noise the pressure and velocity distributions should be suppressed. We describe the generation mechanism of axial fan noise, especially turbulent noise, and the estimation of the noise level and characteristics of an axial fan which was carried out using the measured pressure fluctuation due to blade passage.

View full abstract

In order to study the behavior of severe cavitation erosion occurring in a typical centrifugal pump with a high specific speed of N_s = 300 (m, m³/min, rpm), the high-frequency vibrations and noises due to such cavitation erosion are systematically investigated under various test conditions. The present erosion tests are carried out using three-dimensional impellers made of aluminum alloy. In a field where the flow rates as well as NPSH are low, the vibrating accelerations and the acoustic power sharply increase, so that severe erosion takes place in this field. A large difference in flow pattern is clearly observed with respect to the flow rate.

View full abstract

A correlation, which does not involve any empirical constants dependent on refrigerants, is developed for forced convective boiling heat transfer inside a smooth tube. The correlation is based on the model where the heat transfer coefficient of a mixture is lower than that of the equivalent pure fluid due to the mixture effects on nucleate boiling and to the heat transfer resistance in the vapor phase. The predicted heat transfer coefficients agreed with the experimental data with mean deviation of 5.9 and 6.5% for pure refrigerants (R11, R114, R123, R12) and binary mixtures (R11/R114, R123/R134a, R11/R12), respectively.

View full abstract

The impinging jet is a widely used technique for realizing high heat transfer rates between a fluid and a surface. However, the area of high heat transfer rate is limited to near the stagnation point. In this study, the augmentation of heat transfer remote from the stagnation point in the impinging plane jet system by a rod array located near the wall was experimentally investigated. Each square rod in the array was positioned normal to the flow direction and parallel to the flat plate surface. The distance between the nozzle and the flat plate H, and spacing between the rods and the flat plate surface C were changed to find the optimum values. The largest heat transfer augmentation was obtained for C=1mm, H/B=10, where the nozzle width of jet is B. In this case, the heat transfer coefficient averaged in the area of 2B from the stagnation point is about 1.6 times greater compared to that without a rod array.

View full abstract

Fundamental study was carried out concerning the possibility of continuous slurry ice making using D-sorbitol solution flowing in cooled circular glass and stainless-steel tubes. In the present experiment, the supercooling condition of the flowing water solution was released by injecting ice nucleation material such as fine ice particles into the cooled tube. As a result, three types of operating conditions in the pipe, that is supercooling, continuous ice making and ice blockage, were classified. It was clarified that the ice making efficiency was increased with an increase in the nondimensional cooling ratio, and with decreasing of Reynolds number and the concentration of the water solution. The efficiency of continuous ice making in the cooled tube was greater by 2∼5 times than that of continuous ice making outside the tube under the by supercooling condition. A nondimensional correlation equation for the ice making efficiency was derived as a function of some nondimensional parameters.

View full abstract

The reduction characteristics of flow resistance and heat transfer of water solution flow with the surfactant (Cetyltrimethyl-ammonium Bromide) in tubes were investigated experimentally. The flow resistance and heat transfer of Water solution flow with the surfactant were markedly reduced as compared with those of pure water flow. Useful nondimensional correlative equations of flow resistance and heat transfer were derived in terms of various non-dimensional parameters.

View full abstract

The objective of this paper is to clarify how the heat transfer coefficient changes along the flow due to bubbles passing through a narrow vertical rectangular channel (20 mm wide, 2 mm clearance and 450 mm long). The experiments are performed using subcooled water of 80, 60 and 40 K at atmospheric pressure in which the air bubbles are injected into a channel at designated intervals from 0.125 to 1.0 sec and the bubble length is controlled to be equal to 0.03, 0.02 and 0.01 m. The experiments show that the heat transfer coefficients decrease along the flow and then reach constant values beyond a certain distance from the leading edge the heated surface where the flow reaches a fully developed state in both the velocity and the thermal conditions. Under the fully developed conditions, the heat transfer coefficients are predicted well by the existing theoretical analysis in which both the convective term and evaporation on the interface are ignored.

View full abstract

The present work investigates analytically non-Fourier effects in an infinite plate as a finite medium subjected to a periodic heat flux using the hyperbolic heat conduction model. By comparing the results of the temperature profiles with those obtained from the Fourier parabolic heat conduction equation, the transition condition between the 'parabolic' and 'hyperbolic' behavior at the front and rear surfaces is obtained. The phase difference between the front and the rear surface after temperature wave propagation through the medium is also calculated numerically as a function of the relaxation Fourier number.

View full abstract

The flow pattern of liquid in an annular tube rotating around the horizontal axis is vizualized experimentally as the flow simulation of working fluid in a heat pipe. Correlations between volume ratio of heat medium and the critical rotation speed for the rimming or collapsing phenomena are obtained. After that, the heat transfer rate of the heat pipes are measured and it is shown that the transitional rotation speeds for the heat transfer agree well with those for the flow pattern.

View full abstract

Space- and time-resolved information, as well as macro- and microscopic visualizations, is essential for detailed study of mixing, reaction and heat transfer mechanisms in a turbulent diffusion flame. However, turbulent diffusion flame structures have been studied almost solely by time-resolved single-point measurement due to the restriction of diagnostics. The time-dependent spatial structures of turbulent diffusion flames have often been masked by the statistical analyses. In this study, time-dependent spatial mixing processes in turbulent diffusion flame are visualized by instantaneous flame luminosity and plane Mie scattering images by an ultrahigh-speed video system (Fastcam). It is found that the autosimilarity of the instantaneous turbulent diffusion flame structure has a fractal dimension defined by fractal analysis which appears corresponding to the fractal analysis of the combustion region. Macroscopic movements of the flame region are classified in terms of the instantaneous variations of flame shape.

View full abstract

A numerical study for self-ignition of single fuel droplet (n-heptane) in a high-temperature environment was performed in this article. The numerical simulations include detailed chemistry in terms of a set of elementary reactions using up to 256 elementary reactions, as well as a multispecies transport model. The ignition time was compared to the experimental results and it was shown that both were found to be in good agreement. Further, the computed results (temperature distribution, ignition time, heat release rate and so on) were compared with the results which were obtained under assumption of single-step reaction model and the validity of single-step reaction model for ignition process was examined. It was calculated that the single-step reaction model does not give appropriate results for ignition time and initial heat release distributions. In case of a small droplet, heat release rate in early time becomes much larger than that for large droplet size, thereby retarding ignition process significantly. This may cause no ignitable condition, which we call the ignitable limit. Then, the mechanism of ignition and ignitable limit were also discussed.

View full abstract

The present study focuses on the problem of a liquid jet injected transversely into a subsonic air stream. Droplet sizes and droplet velocities were measured, and empirical equations of mean droplet size and mean droplet velocity profile were deduced. On the basis of the observation, the spray formation mechanism is classified into two mechanisms : one due to the instability of the liquid jet surface and the other due to the instability of the liquid jet itself. The maximum size of the mean droplet size profile in the liquid injection direction decreases with increasing air velocity and decreasing momentum ratio of the liquid and the air. The location in the liquid injection direction where the mean droplet velocity becomes minimum is coincident with the location where the mean droplet size becomes maximum. There is a strong correlation between the droplet size and droplet velocity, and larger particles have smaller velocities.

View full abstract

The mass flux distribution of spray formed by liquid jet disintegration in subsonic air streams was measured experimentally, and the empirical equation of the mass flux distribution was deduced. The configuration of the upper profile of the spray mass flux in the liquid injection direction above the peak of the profile is different with that of the lower profile. With increasing the momentum ratio of the liquid and the air, the location in the liquid injection direction where the mass flux becomes maximum is elevated, and the maximum mass flux decreases. With increasing momentum ratio, the number of spray droplets formed at the liquid jet surface before the liquid jet breakup point increases, and contour lines of the spray mass flux show a kidney-shaped configuration near the liquid jet breakup point. Downstream of the liquid jet breakup point the maximum mass flux decreases, and the mass flux profile becomes flatter.

View full abstract

The influence of tumble intensity on the three-dimensional flow-field structure was analyzed numerically using the STAR-CD code. It was found that when the tumble intensity is extremely high, because the velocity component in the direction of the axis of tumbling air motion is small, the flow-field structure can be treated as two-dimensional. When the tumble intensity is moderate or weak, however, various flows in the direction of the tumble axis can be observed. Therefore, such a flow-field structure should be treated as three-dimensional. A new method to measure the three instantaneous velocity components from a two-dimensional photograph was developed. By using three laser sheets composed of two continuous and one pulse laser light source with different colors radiated for preset durations and timings, velocity in the direction of the laser sheet depths can be derived. Employing this method, the flow-field structure at 15°BTDC was analyzed. It was confirmed that the distortion of tumble induces horizontal vortices in addition to the vertical vortex.

View full abstract

In order to improve fuel economy and reduce emissions in gasoline engines, fine fuel atomization is required. This report shows how the fuel atomization of an injector affects the engine performance. This injector is referred to as an upstream-swirl-type injector, because it consists of a lightweight ball valve and a fuel swirling element (called a swirler) that induces a swirl at the upstream of the metering orifice. Atmospheric conditions such as low air pressure and low temperature did not have any significant effect upon the shape of the fuel spray. A small spray angle and a mean droplet diameter of less than 100μm (Rosin-Rammler) is obtained. The proposed fuel injector shows better acceleration response and starting performance below-30°C in engine tests using passenger car engines compared with the conventional Hitachi injector.

View full abstract

To improve the performance of engines using electronically controlled gasoline injection, the fuel control range must be increased and the injection precision must be improved. This report describes the injection characteristics (fuel quantity vs input pulse width) of the previously developed up-stream-swirl-type injector. Fuel quantity of this injector tends to increase in the small injection area because of the remaining non-swirled fuel. In order to extend the effective fuel control range, it is necessary to linearize this little injection area. It became clear that to reduce the amount of bump and jump using lightweight valve assembly and decrease of the quantity of the remaining non-swirled fuel have significant effects.

View full abstract

Recently, the four-valve engine has become an increase of common type of gasoline engine. However, in such engines, the fuel supply is delayed because of the central wall of the intake port. Thus, the characteristics of a spray through an injector must be improved. This report presents the analysis on the fuel flow and experimental results on the characteristics of the spray injected through an upstream-swirl-type injector which was previously developed by authors. Installation of an adapter for splitting, the fuel spray below the metering orifice produces a two-stream spray. This adapter is a single-port type that has parallel walls with expansion chambers on the left and right (spectacle shape). Spray separation can be monitored by adjusting the position of these two parts. The two-stream injector atomizes and distributes fuel very well. It also improves starting and acceleration, as well as overall engine performance.

View full abstract

Thermal conductivity and thermal diffusivity of combustion chamber deposits in an s. i. engine have been measured to provide the data for analyzing heat transfer which is reduced by the deposits. The deposits were built-up on the surface thermocouples under different engine speeds, loads and equivalence ratios. After removing the surface thermocouples with deposits from the engine, the deposit surface temperatures were measured by an infrared technique in a test cavity installed in a motored engine and at the same time the deposit-wall interface temperatures were detected with the surface thermocouple. The thickness of the deposits was measured with a laser displacement meter. These data were used for determining thermal conductivity. Thermal diffusivity was obtained from penetration time measurements for a laser pulse emitted to the deposit surface. The results show that thermal conductivity is almost unchanged for different engine speeds, loads and equivalence ratios, while thermal diffusivity is influenced by the equivalence ratio.

View full abstract