The behavior of a deformable body with a viscoelasitc membrane in fluid flows is numerically investigated by the two-phase lattice Boltzmann method. In the calculations, the internal fluid of the body is distinguished from the surrounding fluid so that the viscosities of these fluids can be separately specified. Using this method, the deformation of the body under a shear flow is simulated by changing the viscosity ratio η, which is defined as the ratio of the viscosity of the internal fluid to that of the surrounding fluid. It is found that as the viscosity ratio increases, the deformability of the body remains almost constant for 0.1 ≤ η ≤ 1, whereas it decreases linearly for 1 < η ≤ 10. In addition, the behavior of the body in a square pipe flow is simulated for various viscosity ratios. The velocity near the centerline of the pipe decreases with increasing the viscosity ratio, and thus the flow rate becomes reduced. In the case of lower viscosity ratio (0.1 ≤ η ≤ 1), the body becomes a parachute-like shape and approaches toward the centerline of the pipe; that is, axial accumulation is observed. In the case of higher viscosity ratio (e.g., η = 6), the body forms into a slipper-like shape, migrating laterally toward a certain equilibrium position between the wall and the centerline of the pipe.

View full abstract

A new numerical method based on the ghost fluid method was developed for compressible two-phase flows; the idea of adaptive mesh refinement with multigrid was implemented in the method. In the method, interpolation techniques between multiple grids near interfaces were also proposed. The present techniques are effective in diminishing the numerical instability caused by the discontinuity of physical variables across the interfaces. The bubble collapse induced by the interaction of an incident shock with a gas bubble in water was simulated with the present multigrid ghost fluid method. We have succeeded in capturing the fine interface and vortex structure during the collapsing and rebounding stage. The mass conservation is improved with the adaptive mesh refinement combined with the hybrid particle level set method. Also, the interaction of an incident shock wave with a gas bubble near a deformable boundary was simulated successfully with the method in which the motions of three phases, i.e., the gas inside the bubble, the ambient liquid, and the wall material, are taken into consideration.

View full abstract

Aerodynamic noise from bluff bodies is well reduced when they are covered with porous materials. In this study, presence of porous materials is modeled by momentum loss in the equation of fluid motion, and mathematical formulae for evaluating aerodynamic noise based on the theory of vortex sound are modified in order to include the effect. Flow around the cylinder covered with porous materials is calculated by LES techniques. When the inertial resistance factor C₂ is set to 200[m^-1], no vortex is shed periodically into the wake any longer and a region of zero-velocities is spread behind the cylinder, which forms stable shear layers. Time variations of the force exerted on the cylinder and the total momentum loss in the whole porous areas are also reduced at C₂=200[m^-1]. Flow through porous materials induces velocities in circumferential direction at the outer edge of the materials which weaken vortices of shear layers, and also induces velocities in radial direction which widen the space between two shear layers behind the model. Numerical analysis predicts the reduction of aerodynamic noise in the case of C₂=200[m^-1], which agrees qualitatively with experimental results. Although strong sound sources are formed around separation points in the case of the bare cylinder, stable flow around the cylinder with porous materials seems to suppress production of sound sources.

View full abstract

In order to clarify the relations between vehicle motions and flow fields, LES (Large Eddy Simulation) is conducted for flows around a simplified car model in pitch motions, where an improved discretization scheme in the ALE (Arbitrary Lagrangian-Eulerian) coordinates is employed. The analysis model is the Ahmed model with a rear slant angle of 12.5 degrees. It is forced to oscillate sinusoidally with forcing amplitudes of 0.88 and 3 degrees at two different frequencies of 4 and 8 Hz. The fluctuations of the lift coefficient (C_Z) are phase-lead to the pitch angle (Δθ), while those of the pitching moment coefficient (C_MY) are phase-lag to Δθ. Especially in the cases of the larger amplitude of 3 degrees, the Lissajous curves of C_MY form characteristic teardrop shapes. In addition to the three aerodynamic effects observed in the heave motion, such as flow contraction, relative change of inflow angle, and fluid added mass effect, pressure on the upper and under surfaces of the model fluctuates due to the change of incident angle. It is also revealed that the streamwise vortices which induce flow from the underside to the lateral side, are formed close to the side surfaces of the model when the model is in the pitch-up position. These vortices are closely related to the aforementioned teardrop-shaped Lissajous curves of C_MY.

View full abstract

The streamwise interface of an isolated turbulent region in a laminar flat plate boundary layer is investigated by performing a wind tunnel experiment. In the experiment, bimorph-type piezoceramic actuators are used to generate a trapezoidal-shaped turbulent region that has a wide interface in the spanwise direction. With the use of the piezoceramic actuators, the interference of the spanwise interface with the streamwise interface of the turbulent region is eliminated. Therefore, the streamwise interface can be considered as an independent interface. The experimental results show that the rms value of the velocity fluctuation and the turbulent fluctuation rapidly changes at the leading edge of the turbulent region. On the other hand, the rms value of the velocity fluctuation remains high after the turbulent fluctuation disappears at the trailing edge of the turbulent region. It is also shown that the traveling speed of the trailing edge of the turbulent region is faster than the ensemble-averaged streamwise velocity at the trailing edge of the turbulent region near the wall (η≤1.3). In addition, the velocity profile at the trailing edge of the turbulent region indicates the existence of the acceleration of the turbulent fluid near the wall. From the above results, it is found that the inverse transition occurs at the trailing edge of the turbulent region near the wall due to the acceleration of the turbulent fluid.

View full abstract

To reveal the nature of turbulence in a lung, the measurements of turbulence in the oscillatory flow in realistic model human central airways were made by particle image velocimetry (PIV). The transparent silicon model of multi-branching airways was fabricated from X-ray CT images by rapid prototyping. The multi-branching airways comprise trachea, right and left bronchi and airway diameters range from 14 to 2 mm. Experiments were performed for the Reynolds number from 1200 to 2200 and for the Womersley number from 1.9 to 2.3 in the trachea. The spatial and temporal variations of turbulent intensity were strongly dependent on the airway geometry and the phase of oscillatory flow. The expiratory flow generates strong turbulence which explosively occurs in the entire cross section especially in the right bronchi, whereas the inspiratory flow generates relatively weak turbulence near the airway wall.

View full abstract

Swirling flow behavior between two parallel disk shape plates was experimentally investigated with the aid of a particle image velocimetry (PIV). The experiment was performed at low Reynolds numbers (Re < 100) to simulate the practical operation in a disk shape planar-type solid oxide fuel cell (SOFC). In the channel installed radial-type current collectors, unfavorable flow deceleration occurred toward downstream direction. Since this result suggested the necessity of improvement of flow uniformity, we designed a new channel with circle involute shape current collectors. In the new involute-type channel, a swirling flow was generated and its velocity was kept at nearly constant value toward the channel exit. This trend was observed regardless of flow rates, and hence flow uniformity was achieved over the wide range of Reynolds numbers. This is because a flow passage consisting of two adjacent involute shape current collectors functions as a constat-area channel due to the geometrical property of the circle involute. Furthermore, an estimation of a fluid motion in the involute-type channel was carried out by using steady state Euler's equation of motion. We confirmed that the velocity component in the flow direction was dominant compared with that in the other direction and played primary role to maintain a swirling motion through the centripetal acceleration term.

View full abstract

The theoretical calculation of a two-stage high-velocity oxy-fuel thermal spray gun was conducted for the purpose of predicting the stagnant pressure in the mixing chamber (MC). The theoretical results were compared with the experimental ones obtained by the cold flow and actual spray gun tests. The effect of the diffuser shape installed between the combustion chamber (CC) and the MC was also investigated in the cold flow test. It was clarified in this study that 1) the theoretical calculation shows that in order to obtain a larger pressure in the MC the diameter rations of d_3t/d_1t and d_1e/d_1t should be chosen to satisfy Eq.(11), to keep the diffuser throat unchoked, 2) the theoretical results of p₀₃/p₀₁ agree with the results of cold flow test within an accuracy of 6%, 3) the theoretical results of p₀₁ agree with the results of actual two-stage HVOF gun test within an accuracy of 4%, demonstrating the validity of the present theoretical calculation.

View full abstract

The influence of different parameters in the design of multi-blade fan , such as the ratio inner diameter of outer diameter, D₁/D₂ , number of blades, Z, inclination angle,α, outlet angle,β_d2 , and inlet angle,β_d1 , were examined in order to study relation between the flow pattern and noise level. In the present report, we investigated the fan performance and the noise characteristic experimentally and compared with simplified two-dimensional CFD analysis. The flow pattern passing the blades under the different geometrical conditions was examined and discussed in terms of fan performance and noise level. The results shows that high efficiency and low noise were attained in range of D₁/D₂=0.8-0.84 , Z=35-45 , α=30°-40°, and β_d2=155°-165°. For these conditions, it was found that the pressure loss and noise generation in the flow between blades were attributed to the separation of the flow which was induced by the circulated vortex at the suction side, reattachment of the separated flow, and shear flow at the trailing edge. It was also confirmed that a large inner to outer diameter ratio gives a defected performance and big noise generation because of no reattachment of the separated flow. Also, the increasing of blade number resulted in the increasing of high frequency noise due to the vortices generation in shear layer in spite of rectification of passing flow.

View full abstract

Linear stability analyses and numerical computations have been carried out for a magnetic Rayleigh-Bénard convection of air in a coaxial double cylinder in the absence of gravity field. An inner cylinder is heated and an outer cylinder is cooled both isothermally. The air filled within the gap of double cylinder is unstable under certain strength of the applied toroidal magnetic field since the heated air is repelled radially outward due to the magnetic buoyancy force. The stability analysis quantitatively predicts the critical values for the onset of thermal convection for various values of the radius ratio.

View full abstract

Heat transfer enhancement for a laminar natural-convection boundary layer along a vertical flat plate by inserting short flat plates set in array at fixed intervals (split heat transfer promoter) has been numerically analyzed to comprehend the mechanism originating the increase in heat transfer rates. The calculated results impressively reproduce visualized fluid motions obtained in the previous experiment, and it is again confirmed that the remarkable heat transfer enhancement is due to the synergetic effect of low temperature fluid flow detouring the inclined plates and the longitudinal vortexes generating at the edge faces of inclined plates. For the heated surface of 0.3 m height, the use of 7 column split heat transfer promoters conduces to an increase in average heat transfer rate of approximately 50 % compared with that without promoter.

View full abstract

The electrospray characteristics depend on liquid flow rate, applied voltage, electrode configuration and liquid properties such as electrical conductivity, surface tension and viscosity. This study experimentally investigated electrospray characteristics of binary liquid mixtures. The tested liquid was ethanol, which has relatively high electrical conductivity and low heating value as a liquid fuel, n-heptane, which has relatively low electrical conductivity and high heating value, and their mixtures. The electrical conductivity of the liquid mixture decreased rapidly with increasing heptane volume fraction for heptane volume fraction higher than 70 %. The electrospray characteristics for mixtures with heptane volume fraction less than 70 % were different from those with heptane volume fraction greater than 70 %. By increasing the applied voltage, the electrospray mode changed in the order as spindle, cone-jet and multijet below the heptane volume fraction of 70 %. For the heptane volume fraction of 90 %, however, the electrospray mode changed in the order as spindle, cone-jet, asymmetry cone-jet and plused jet. The onset voltage, over which the cone-jet mode appears, was decreased with increasing heptane volume fraction while heptane volume fraction was less than 70 %.

View full abstract

In this paper, the formula for obtaining the absorption horsepower of Moulinet was redrawn, and the physical meaning of the constant in the formula was clarified. Based on this study, it was verified experimentally about the designing method of Moulinet and the calibration method of Moulinet after manufacture. Consequently, the followings were clarified; (1) if the propeller power coefficient was taken to the proportionality constant, the absorption horsepower of Moulinet was proportional to the cube of the revolution, and the fifth power of the Moulinet diameter. (2) if it was Moulinet which carried out the similarity design geometrically with the standard size of Aviation Technical Research Center's type-6 Moulinet, the proportionality constant C₁ given in the reference could be used, and the absorption horsepower of the Moulinet was proportional to the cube of the revolution, the cube of the Moulinet diameter, and the side projection area of the Moulinet. (3) the proportionality constant C₁ was proportional to the propeller power coefficient C_P.

View full abstract

The membrane electrode assembly (MEA) durability test protocols proposed by FCCJ, USFCC and DOE were compared in their deterioration behavior of performances and materials. These MEA durability tests for polymer electrolyte fuel cells were performed using JARI's standard single cell. In the carbon support corrosion test protocols, carbon decomposition (oxidation), collapsing of the porous structure of the carbon support, and agglomeration and dropout of the platinum (Pt) catalyst were observed. A proper fast evaluation can be done by the FCCJ protocol because of the highest applied voltage. In the Pt catalyst stability test protocols, both Pt particle growth and also carbon oxidation were observed in the Pt/C catalyst used in this study. It was found that the carbon oxidation rate was the lowest in the DOE protocol. In the membrane durability test protocols, decomposed fluorine F- and thinning of membrane were observed. Tests results suggest that the FCCJ protocol (using air, without pressure difference between anode and cathode) was the best at early stage evaluation.

View full abstract

Cold start characteristics of a polymer electrolyte fuel cell were investigated experimentally, and microscopic observations were conducted to clarify the freezing mechanism in the cell. The results shows that the freezing mechanisms are classified into two types: freezing in catalyst layer at very low temperature like -20°C, and freezing of supercooled water at the interface between cathode catalyst layer and micro porous layer at near 0°C like -10°C. The amount of produced water in each the processes is related to the initial wet condition of the membrane because there is a period of back diffusion of produced water into the membrane. It is also shown that the performance of a subsequent normal temperature operation at 30°C after the shutdown in the cold start is temporarily deteriorated after the freezing at -10°C, but not after the freezing at -20°C. The ice formed at the interface between the catalyst and the micro porous layers is estimated to cause the temporal deterioration, and the function of micro porous layer on the gas diffusion layer is also discussed.

View full abstract

Comprehension of the burning velocity for micro-scale flames is inevitable in the improved design of micro-combustors for miniaturized power supplies. The present study is performed to examine experimentally the burning velocity characteristics of hydrogen-premixed micro-scale spherical laminar flames in the range of flame radius r_f < 5 mm, and also macro-scale laminar flames with r_f > 7 mm for comparison. The mixtures have nearly the laminar burning velocity S_L0 at so-called unstretched flames with different equivalence ratios (φ=0.3-1.2). In this experiment, the values of the S_L0 are fixed at approximately 15, 25 and 35 cm/s, in order to examine the influence of characteristic chemical reaction time on micro-scale spherical laminar flames. The radius and the burning velocity of micro-scale flames are obtained by using sequential schlieren images recorded under appropriate ignition conditions. The results show that the burning velocities of micro-scale flames with φ=0.3 and 0.5 or 1.2 have a tendency to decrease or increase with increasing r_f and approach that of macro-scale flames, but such a trend can not be seen for φ0.7 and 0.9 micro-scale flames, irrespective of S_L0. It is also found that the optimum size and Karlovitz number to improve the burning velocity are existed for φ0.7 and 0.9 micro-scale flames, irrespective of S_L0.

View full abstract

The effect of the electrode on the flame kernel initiation in the spark ignited methane/air mixtures has been numerically investigated. The electrode gap was thought to have strong effects on the flame kernel initiation and was varied as a parameter. The equivalence ratio of the mixture was also varied. A detail reaction mechanism of methane and a detail transport mechanism including thermal diffusion were taken into account in the numerical model. The effect of the heat loss to the electrode was investigated for the several condition of the electrode surface. Results indicated that the heat loss to the electrode for the isothermal condition of the electrode surface was not different from that for the thermal condition including the heat transfer to the electrode and the heat conduction in the electrode although the flame kernel initiation for the adiabatic surface of the electrode gap was rather different from them. The flame kernel initiation in the stoichiometric mixture was enhanced with increase in the electrode gap for the same ignition energy. The OH production rate in the lean mixtures was smaller than that in the stoichiometric mixture for the same ignition energy. Even in the lean mixture, OH production rate became much greater when the electrode gap was broadened only from 3 to 4 mm.

View full abstract

The effect of NO doping on high temperature oxidation reaction was investigated to have insights into the control of HCCI combustion. Numerical analysis based on GRI-Mech methane oxidation scheme showed that the doped NO inhibited methane oxidation. The engine experiments were conducted to verify the calculation results using a single cylinder HCCI test engine. The experimental results showed that the doped NO enhanced the oxidation reaction, which were completely contrary to the calculation results. The largest cause for the error was in CH₃ oxidation path in GRI-Mech.

View full abstract