Recent development for the direct numerical simulation of dispersed multiphase flows is reviewed. A current trend is to use the fixed Cartesian mesh, because it is suited to deal with multiple solids, bubbles and/or droplets. Features of typical methods for gas-liquid interfaces and solid-fluid boundaries are discussed. For gas-liquid interfaces, existing methods are categorized into fitting, tracking, capturing and transportation methods in order to clarify their characteristics. For deformable solid objects moving in a fluid, on the other hand, schemes based on the immersed-boundary method are highlighted. Our prime interest is in the conservation properties of the algorithms. From such a viewpoint, we propose the combination of volume-of-fluid (VOF) method and the velocity-field overridden by immersed solid (VOIS) method for flows laden by multiple bubbles and particles.

View full abstract

A single-physics CFD simulation has been established now, and it is daily used in design and analysis processes of various machines. Therefore, a multi-physics CFD simulation would be a key technology to be researched and developed in near future. We can easily find a number of multiphysics phenomena in nature and machines, but we cannot simulate most of them. That is, there still remain many difficulties in a multi-physics CFD. For example, the modeling of physics except for heat and fluid flow has not been developed yet. In the present article, I briefly explain the state of the art of a multi-physics CFD simulation, focusing on the coupling method between flow and other physics, and then describe our researches on (1) sand transfer in a desert and its suppression with a wind-breaking fence and (2) particle deposition phenomena on the turbine vane in a jet engine. In both simulations, a weak coupling method was employed, taking into account the time scales of the physics embedded in the phenomena. Using the computational results, the characteristics of each phenomenon were discussed and some useful insights for improving the multi-physics simulation were obtained.

View full abstract

Flow simulation is growing up as Computational Fluid Dynamics to an important field of engineering and research of fluid dynamics. In order to contribution to society demands, it has been developed for various problems including complex phenomena such as turbulence, sound, multiphase materials, reaction and so on. The trends of these development are discussed by the categories of society demands on energy, material and human and their specific research topics on Computational fluid dynamics.

View full abstract

In this paper, we investigated an advantage of the finite difference lattice Boltzmann method (FDLBM) to direct numerical simulations of sound propagation including scattering by a body. In order to incorporate, monopole sound source into the flow, a new FDLB formulation was proposed using the Chapman-Enskog expansion. We calculated the acoustic scattering problem around a circular cylinder using the FDLBM discretized by non-compact finite difference scheme which is faster than compact finite difference scheme. The results agreed with both the exact solution and the accurate calculation by the macroscopic equations based on the discretizations of compact finite difference scheme. As a result, the FDLBM was proved to have a merit to a calculation of sound propagation. Also, we estimated error of the FDLBM in the acoustic pulse problem.

View full abstract

In this paper, the seamless virtual boundary method is applied to the incompressible flow analysis with heat transfer. In the seamless virtual boundary method, the forcing term is added not only on the grid points near the boundary but also on the grid points inside the boundary, in order to satisfy the velocity conditions on the virtual boundary points. Then, the smooth physical quantities can be obtained. For the heat transfer, the forcing term satisfied the temperature condition has to be added to the energy equation. First, we validate the present approach for the natural convection in square cavity. The results show that the present solutions are in good agreement with reference solution. Next, we try to apply the present approach to the flow around a heated circular cylinder. Then it is concluded that the present seamless virtual boundary method is applicable to the incompressible flow analysis with heat transfer successfully.

View full abstract

Wake structure of permeable object is one of great concerns for wind engineering, chemical processing, and fluid machineries. It produces the drag greater than the same-sized solid object due to friction dominant nature even at Reynolds numbers higher than 10^3. In order to characterize the wake structure in such a condition, we measure the flow field around the permeable cylinder made of wire mesh and annularly arrayed thin cylinders. The parametric study presents a spatially delayed amplification of turbulence due to vortex shedding in two frequencies compound in the wake. We also measure the flow for a rotating permeable cylinder to find the interaction between circulating elements and macroscopic behavior of flow. The result indicates a curtain effect of the elements to expand the momentum loss region behind the rotating cylinder.

View full abstract

Aiming at elucidating the relationship between injector internal flow, especially injection velocity profiles, and atomization characteristics of liquid sheet, experimental measurement, numerical and theoretical analyses were carried out. Liquid space distribution produced by atomization of axisymmetric sheet, which was created by collision of two water jets in opposite direction, was experimentally obtained utilizing paternator. The amplitude of the liquid sheet oscillation was developed based on linear stability analysis. The numerical results of atomization process through Kelvin-Helmholtz type instability showed qualitative resemblance with corresponding experimental results and theoretical analyses. It was clarified that non-uniform injection velocity profile resulted in velocity distribution with inflection point inside the sheet, thus the sheet became unstable and enhanced atomization. The effect of injection velocity profile on atomization at impingement type injector was also clearly represented.

View full abstract

Diameters of micro-bubbles are to range from about one to several hundreds μm, and therefore it is difficult to measure the correct diameter distribution using a single measurement method. In this study, diameters of bubbles generated by a pressurized dissolution method are measured by using phase Doppler anemometry (PDA) and an image processing method, which is based on the Sobel filter and Hough transform. The diameter distribution and the Sauter mean diameter of micro bubbles are evaluated based on the diameters measured by both methods. Experiments are conducted for several mass flow rates of dissolved gas and of air bubbles entrained in the upstream of the decompression nozzle to examine effects of the entrained bubbles on bubble diameter. The experimental result indicates that diameter distribution of micro bubbles can be accurately measured for a wide range of diameter by using PDA and the image processing method. The mean diameter of micro-bubbles generated by gasification of dissolved gas is smaller than that generated by breakup of air bubbles entrained in the upstream of the decompression nozzle. As the mass flow rate of the entrainment of air bubbles in the upstream of the decompression nozzle increases, the mean bubble diameter increases and the bubble number density decreases.

View full abstract

A numerical method based on the preconditioned method coupled with PROPATH is applied to mixing flows of supercritical water and water liquid in a T-shape channel which is used as a reactor of supercritical hydrothermal synthesis. PROPATH has accurate thermophysical properties of gas, liquid and supercrtical-fluids. Since the thermophysical properties such as density and viscousity change anomalously near the critical point, the mixing flow of supercritical-fluid and liquid is quite different from that of gases or liquids. The calculation ignoring several thermophysical property changes indicates that accurate thermophysical properties are certainly required for evaluating supercritical-fluid flows with phase change. The prediction of buoyancy effecrt to the flows is also quite important to get the real flow fields.

View full abstract

With the progress of human activities in space, the occasion to handle liquids in no-uniform acceleration or low-gravity is now growing. On the launch vehicles with liquid propulsion system, the dynamic acceleration during its powered ascent or ballistic flight makes it very difficult to control the position of propellants in the tanks. For the establishment of the technology for the management of liquid propellant in space vehicles, a numerical method, called 'CIP-LSM' (CIP based Level Set & MARS), was developed to simulate three-dimensional free-surface flows under various gravity conditions, which has been applied to clarify the dynamic behavior of liquid propellant in the tanks of launch vehicles.

View full abstract

This paper investigates a thermal-fluid dynamics of CAMUI (Cascaded Multistage Impinging-jet) type hybrid rocket developed in Hokkaido University by using a large eddy simulation of turbulence. The performance of the hybrid rocket is sensitive to the changing shape of its chamber. To clarify this effects, numerical simulations were conducted using measured shapes. The results show the flow structures such as impinging fountain flow depending on the shapes at different burning time. Thease structures generate the particular heat flux distributions on the surface.

View full abstract

Sand erosion is a phenomenon whereby solid particles impinging on a wall cause serious mechanical damage to the wall surface. This phenomenon is a typical gas-solid two-phase turbulent flow and can be considered as a multi-physics problem in which the flow field, particle trajectory, and wall deformation interact. On the other hand, aircraft engines operating in a particulate environment are subjected to the performance and lifetime deterioration due to sand erosion. Especially, the compressor of the aircraft engines is severely damaged. In order to consider sand erosion in design phase, it is important to develop the prediction method and to obtain basic data. In the present study, we apply our three dimensional sand erosion prediction code to a single stage axial flow compressor. The numerical results of the eroded surface geometry and the performance deterioration showed the same tendency as experiments. Moreover, the change of the flow field and the particle trajectories are investigated to clarify the erosion mechanism of the compressor.

View full abstract

The fluid-acoustic interactions and the role of the acoustic resonance are clarified for the fluid-resonant oscillations in turbulent cavity flows by directly solving the compressible Navier-Stokes equations. The upstream boundary layer is turbulent, and the freestream Mach number is 0.15 or 0.3. The computational results are validated by experiments done in a low-noise wind tunnel. The depth-to length ratio of the cavity is varied from 0.9 to 2.5 and the results are compared with those of the fluid-dynamic oscillations, which take place in a shallow cavity with the depth-to-length ratio of 0.5. It is clarified that the mechanisms of the formation of large-scale vortices and the radiation of acoustic waves are essentially the same as those in the fluid-dynamic oscillations. However, the feedback loop is formed via a standing wave due to the acoustic resonance in the fluid-resonant oscillations, while a standing wave is not generated in the fluid-dynamic oscillations. The acoustic resonance amplifies the acoustic energy radiated by the collision of the vortices along the downstream wall. Moreover, the acoustic feedback becomes more intense due to the acoustic resonance. As a result, the fluid-resonant oscillations occur even at a low Mach number such as 0.15. Moreover, the effects of the cavity depth on the frequency as well as the amplitude of the fluid-resonant oscillations are clarified.

View full abstract

The passage vortex in a centrifugal impeller is generated by the centrifugal force caused by the curvature of impeller passage and the Coriolis force induced by the rotation, and strongly affects the loss generation in the impeller channel. Therefore, the control of passage vortex results in the reduction of loss generation. In the present study, the flow within the rotating curved duct which is considered to be a simplified model for centrifugal impeller channels was analyzed numerically in order to reveal the effects of the inlet pitchwise velocity distribution on the formation of passage vortex. The computed results revealed that the thicker boundary layer on the suction side wall than that on the pressure side at the inlet of the bend suppresses the development of passage vortex and consequently reduces the associated loss generation. Moreover, the results were examined in detail by relating with the flow in a centrifugal impeller.

View full abstract

A wind tunnel experiment was conducted to simulate large-scale turbulent motions in an atmospheric boundary layer (ABL). The large-scale turbulent motions originating from outer-layer disturbance were generated using an active grid installed at the front of the test section and the winglets of the active grid were randomly rotated. The large-scale horizontal motions in the ABL were generated in the wind tunnel using the active grid when appropriate values of the angle and rotation speed were chosen. The validity of the present method using the active grid was confirmed by comparing the integral scale and energy spectra obtained by the experiments with empirical formulas inferred from field observations. This indicates that the present method of using the active grid accurately simulates the turbulent motions in a nearly neutral ABL.

View full abstract

Pressure measurements, laser Doppler velocimetry (LDV) and flow visualization were carried out for a 90° elbow with a curvature that coincides with the pipe inside diameter in order to clalify the structure of a fluid fluctuation occuring in the bend. Results of pressure measurements show that the value of the pressure loss occurring in the elbow coincides with the conventional empirical formula with a turbulent boundary layer. This agreement is applicable to the case that Reynolds number is greater than 50000. Detailed results of LDV measurement for Reynolds number of 50000 is shown. Frequency analyses of axial velocities indicate that two kinds of flow fluctuations exist in the present elbow. One is the fluctuation with approximate Strohal number (based on a pipe inner diameter and a bulk velocity) of 1.0 occurring in a shear flow region. The other is with approximate Strouhal number of 0.5 occurring in the separated and its downstream regions. These fluctuations were also observed by flow visualization.

View full abstract

A valveless micropump has been developed with a diffuser/nozzle shaped channel and a variable volume actuator which produces an oscillating flow. The pressure-loss in a nozzle is lower than that in a diffuser, therefore one-way flow is considered in the nozzle direction. Frequency characteristics were determined. Simplified 1-dimentional analysis was done for unsteady operation of the pump by considering the channel geometries and pressure loss coefficients based on unsteady Bernoulli's equation under the assumption of quasi steady flow. 3-dimentional calculations were done using the commercial CFD. The calculated pump characteristics agreed with the measured ones for lower drive frequency. The measured and calculated results were rearranged using dimensionless numbers considered the direction, and compared the pump characteristics to understand the physical mechanisms of the micropump.

View full abstract

This study is intended to develop the waste heat recovery generation system using steam-Rankine cycle for co-generation system with a small capacity prime mover in the range of about 500kW or less. In our past reports, a scroll expander has attracted attention as a power recovery device which has high efficiency potential even in the case of small power capacity, and the performance characteristics of it are investigated through the experiments. Then, it was confirmed that the leakage loss in the beginning of expansion process and the thermal loss are caused the performance of the scroll expander to decrease. Therefore, in this paper, the shape of the orbiting scroll was redesigned to prevent the leakage, and also the lubricant oil supplied to the expander was heated up to the same temperature of steam as working fluid to reduce the thermal loss. As a result, the expander performance is improved by the reduction of the leakage at the beginning of expansion process, and the expander efficiency reaches up to 59.0% at the maximum in the experimental conditions that the suction pressure of working fluid is 590kPa and the revolution speed is 1850rpm.

View full abstract

The engine performance with three stroke/bore ratios in a spark ignition gas engine was investigated in experiments and with CFD engine analysis. The thermal efficiency improves significantly with the increase in stroke/bore ratio from 1.0 to 1.5 due to reductions in cooling loss and improvements in the degree of constant volume combustion. With further increase in stroke/bore ratio from 1.5 to 2.0, the maximum IMEP and the degree of improvement in thermal efficiency decrease. The experimental equations for the heat transfer co-efficient with a mean piston speed including Woschni's equation do not express the change in heat transfer with the higher stroke/bore ratios.

View full abstract

In this study, the experiments of heat flux measurements were carried out to improve the precision of heat flux sensor. Three kinds of heat flux sensors were made and used in the experiments. These are type-A1, type-A2 and type-B. The type-A1 is the sensor that the material of the sensor body is different from that of the combustion chamber wall, but the material of wire is same as that of the thin film on the sensor surface. The type-A2 is the sensor that the material of sensor body is same as that of the combustion chamber wall, and the material of wire is same as that of the thin film on the sensor surface. The type-B is the sensor that the material of sensor body is same as that of the combustion chamber wall, but the material of wire is different from that of the thin film on the sensor surface. In addition, the type-A1 is the only sensor of them that the materials are not magnetic. As the result, the requirement for the heat flux sensor is that the materials of wire and thin film must be the same. Ferromagnetic materials are not suitable to use for moving parts as the piston head of an engine. The material of sensor body does not need to use the same material of the combustion chamber wall.

View full abstract

The heat transfer coefficient on an engine cylinder wall from burnt gas is important information for studies on phenomena of the heat transfer in engine cylinder. It is necessary to obtain the gas temperature near the cylinder wall for calculating the heat transfer coefficient. In this study, the burnt gas thermometer by infrared rays (IR thermometer) is constructed, calibrated and applied to the engine cylinder gas temperature measurements. The calibration experiments of the IR thermometer are carried out using the constant volume vessel as the same gas conditions in the engine cylinder at 30, 47 and 59 degrees of ATDC. The calibration equation between output of IR thermometer and burnt gas temperature is calculated by interpolation from the data of calibration experiments, with consideration of composition and density of the burnt gas in the measuring volume. As the results of experiments using a spark ignition engine, the temperature distribution near the engine cylinder wall measured by the IR thermometer is almost uniform, and the difference from average temperature of burnt gas calculated by the first law of thermodynamics is from 100K to 150K. The reason is thought that the gas flow in the engine cylinder is turbulent flow and the gas is approximately uniform temperature by stirring.

View full abstract

A previous procedure to measure thermal diffusivity and thermal conductivity simultaneously using an inverse solution for one-dimensional unsteady heat conduction can be simplified by improving the inverse solution with moving window clearing the issue related to settlement of time left in previous one. The new procedure does not need any time when the temperature change starts at a sensor position and can simply choose a time duration during which the measured temperature change becomes larger than the required accurate temperature change. The measurement is usually completed within about 1 minute until the temperature rise at the thermocouple position reaches a certain temperature level which makes its error level lower than a couple of percents. This method has merit being independent of surface condition except for the requirement of two or three sensing positions in the material. The accuracy of the estimated values is also similar to the error level of sensor at the position.

View full abstract

A method for measuring the diffusion coefficient of methanol in a polymer electrolyte membrane (PEM) was developed using the NMR method. A circular coil of 0.6mm inside diameter was used as a small NMR sensor. The PEM was inserted in a penetration cell, where methanol solvent is supplied to one side of the PEM and nitrogen gas is supplied to the other side of the PEM. The small NMR sensor was placed on the nitrogen gas side of the PEM. The small NMR sensor detects the NMR signal from the methanol solvent which permeates the PEM. The CH and OH components of the methanol solvent were obtained from the NMR signal by spectral analysis. The methanol concentration in the PEM was determined by the ratio of CH to OH components. The methanol concentration was acquired at intervals of 30s and was measured for 2000s. After 1500 seconds, the methanol concentration in the PEM reaches a steady state. The final methanol concentration was about 20% of the methanol concentration of the solvent. It assumed that the diffusion phenomenon of methanol in a PEM was a one-dimensional transport phenomenon, and the time-dependent change of methanol concentration was analyzed by parameterizing the diffusion coefficient. The diffusion coefficient of methanol in a PEM was determined by comparison with the measurement result of the time change of methanol concentration and the analysis results. The concentration difference diffusion coefficient of methanol in PEM obtained using this method was 3.5^*10^<-10>m^2/s.

View full abstract

The mass transfer characteristics of gas diffusion layer (GDL) are closely related to cell performance in PEFC. In this study, the observation method of internal water distribution in microporous media for GDL by using X-ray radiography was examined to elucidate the enhancement mechanisms of oxygen diffusivity of new type microporous media shown in the previous report. It was shown that the distinction between the areas of void and water in microporous media was possible by using X-ray computer tomography. Further, the water distributions in the stracture of new type microporous media, in which two kinds of porous media having different wettability are arranged alternately (Hybrid type GDL), was visualized and the mechanism of high oxygen diffusivity of Hybrid structure of microporous media was clarified.

View full abstract

Since particulate matters (PM) emissions including soot from diesel cars do harm to our health, a diesel particulate filter (DPF) has been used in the after-treatment of exhaust gas. It is reported that DPF filling with PM causes higher back-pressure and more fuel consumption, and continuously regenerating PM trap system is needed. Then, we have focused on the diesel exhaust gas perfect burning system (DEGPBS) developed by COTEC, Ltd., where soot is trapped and burned by the heater. However, the phenomena in the system are not well understood, because it is difficult to conduct the measurement inside the filter. In this study, we simulated soot combustion and deposition by the lattice Boltzmann method to observe the combustion field in the filter. The inner structure of the filter was obtained by a 3D X-ray CT technique. Results show that the heat and mass transport in DEGPBS are well visualized. It is found that temperature of the filter and oxygen concentration are important factors to burn soot in exhaust gas.

View full abstract

Although some countries use esterificated fuel, which is made out of useless vegetable oil, it makes troubles which are emission of white and black smoke, becoming worse of engine start, and so on. Since biofuels are oxygenated fuels, there are concerns abous production of large amounts of aldehydes, which are oxygenated toxic substances. This paper examines the cause of precipitation of solids in fuel at low temperatures, looks at trends in production of aldehydes from biofuels, compares exhaust characteristics when using 3 types of biofuel and gas oil in a diesel engine, and clarifies what effect constituent fatty acids have on exhaust characteristics. As the percentage of saturated fatty acids increases, the cloud point and freezing point rise, and a certain degree of correlation is evident between the cloud point and freezing point. With methyl oleate, production of aldehydes begins to increase around 400, and the produced amount is small in the low temperature range. With methyl ester fuels, there is an increase in NO_x and a decrease in THC, CO and exhaust smoke density, compared to gas oil.

View full abstract

Diesel combustion model for CFD simulation is established taking account of an auto-ignition process of non-homogeneous mixture. Authors revealed in their previous paper that the non-homogeneity of fuel-air mixture affected more on auto-ignition process such as its ignition delay or combustion duration than the turbulent mixing rate. Based on these results, novel diesel combustion model is proposed in this study. The transport calculation for local variation of fuel-air PDF is introduced and the chemical reaction rate is provided by the local non-homogeneity. Furthermore, this model is applied the RANS based CFD simulation of the spray combustion in a Diesel engine condition. The results show that the combustion process is well described for several engine operations.

View full abstract

This paper investigates the performance, exhaust emissions, and combustion characteristics of a dual fuel diesel engine fueled by CNG (Compressed Natural Gas) as the main fuel. The experiments used two fuels for the ignition: one is OME (Oleate Methyl Ester), a major component of biodiesel, and the other is ordinary gas oil. The CNG supply rate was defined as the heat energy ratio of the supplied CNG to the total heat energy available in the cylinder. The results show that the conditions where operation with CNG/OME is possible are very similar to those of CNG/gas oil. When the CNG supply rate was raised to 75%, the brake thermal efficiency was similar to that of ordinary diesel operation at BMEP=0.65MPa. When the CNG supply rate was higher than 75% ignition became very unstable and the brake thermal efficiency decreased significantly as well as the HC and NO_x emissions increased sharply. The reason for this is considered to be that the appearance of miss fire cycles gave rise to combustion fluctuations.

View full abstract

An experimental study of laminar burning velocity and flame stability lean limit of hydrogen-added, preheated ethanol vapor-air flame is presented. The ethanol vapor-air premixed burner flame is obtained using gas mixture of initial temperature 112℃ and 159℃. Laminar burning velocities at the initial temperature of 159℃ are higher than those of 112℃ in all range of the equivalence ratio. Hydrogen addition decreases the flame stability lean limit of the equivalence ratio from 0.66 to 0.27 and increases the laminar burning velocities, in comparison with the same lower heating value mixtures without hydrogen. Lower heating value decreases from 1.84MJ/kg to 0.99MJ/kg in the ethanol vapor-air mixtures at the lean limit.

View full abstract