Chemical vapor deposition (CVD) process composes a complex system, where chemical reaction and heat and mass transfer interact with each other. And these macro-scale phenomena are deeply related to the micro-scale mechanics. Hence, multi-scale analysis is required to understand these complicated interactions. In the previous report, the practical methodology for the multi-scale analysis of reactive rarefied gas flow was proposed. And as the first step of the analysis, the new pair potential model of SiH₄ molecule was constructed using ab initio molecular orbital calculations. However, the interactions between SiH₄ and other species were not considered in the previous report though SiH₄ is seeded into buffer gas in the most of the processes. Hence, we extend the potential model to SiH₄/H₂ system in this report. First, the potential parameters for Si-H and H-H interaction are determined by the molecular orbital calculations of SiH₄-H₂ pair. Next, the Si-Si parameters are determined from the SiH₄-SiH₄ pair. In this manner, the new pair potential model that has the common parameter to SiH₄-H₂ and SiH₄-SiH₄ interaction is constructed. The validity of the model is evaluated from the calculation of SiH₄ viscosity by the molecular dynamics simulations. The result shows the good agreement between the experiments and the model.

View full abstract

Chemical vapor deposition (CVD) process composes a complex system, where chemical reaction and heat and mass transfer interact with each other. And these macro-scale phenomena are deeply related to the micro-scale mechanics. Hence, multi-scale analysis is required to understand these complicated interactions. In the previous report, the new pair potential model for SiH₄-H₂ and SiH₄-SiH₄ interaction was successfully determined using ab initio molecular orbital calculations. And the validity of the model was evaluated by calculating the viscosity of SiH₄ using the molecular dynamics simulations. The result shows the good agreement between the experiments and the numerical calculations. As the next step of the multi-scale analysis, the total cross section model and the scattering angle model for SiH₄/H₂ system are constructed in this report. Many binary collisions for SiH₄-H₂ and SiH₄-SiH₄ are calculated by classical trajectory (CT) calculations with the various initial conditions. Through the statistical analysis, the total collision cross section model is constructed as a function of the relative translational energy only. And the probabilistic scattering model for polyatomic molecules is developed by modifying the classical scattering theory. At last, the validity of these models is confirmed through the comparison of the differential cross section.

View full abstract

Numerical simulation of nematic liquid crystalline flow between rotating parallel disks has been performed using the Leslie-Ericksen continuum theory. It is assumed that the director lies in the shear plane and that the velocity profile is linear in the gap direction of the two disks. For a tumbling-type liquid crystal both the two-dimensional analysis and the Carlsson and Skarp model, in which molecular field elasticity in the radial direction is ignored, can predict step-like changes of the director orientation angles in the radial direction at high Ericksen numbers. The orientation angles are predicted to be discontinuous at tumbling regions for the Carlsson and Skarp model, whereas they are continuous for the two-dimensional analysis. A peculiar behavior observed in a light intensity experiment reported by Carlsson et al. That dark rings move from the disk periphery to its center is satisfactorily simulated by the two-dimensional analysis.

View full abstract

Numerical analysis has been performed for three-dimensional and time-depending vortex breakdown in a confined flow generated by a rotating disk. Vortex breakdown in swirling flows has been the subject of much attention sine it was first recognized in the tip of deltawinged aircraft. Under certain conditions vortex flows undergo sudden structural changes near their rotation axis, called vortex breakdown, which are characterized by the existence of free stagnation point upstream of a region with reversed axial flow. In the numerical calculation, boundary fitted coordinate system has been used for this flow in order to predict asymmetry vortex breakdown. Calculated results of mean velocity are compared with the experimental data in order to examine the validity of the presented numerical method. It has been pointed out as a result of comparison with experimental data that the present method can reproduce vortex breakdown reasonably and predict well mean velocity profile except for the mean velocity profile in the vicinity of rotating lid. Besides, additional calculation has been done to examine the asymmetry vortex breakdown by inclining rotating axis. Calculated results of steam function profile suggest that a pair of vortex generated in symmetric confined flow gradually transforms into asymmetry vortex with inclining rotating axis. By contrast, pressure distribution, that drives the vortex breakdown and shows the existence of adverse pressure gradient followed by positive pressure gradient, is not affected by inclining rotating axis. These results have implied that flow visualization method using dyes is not suitable for clarifying the mechanism of vortex breakdown, because pressure distribution, that is driving force for vortex breakdown, is not influenced on inclining rotating axis.

View full abstract

Numerical analysis has been performed for three-dimensional developing turbulent flow in a square duct with rib-roughened walls by using an algebraic Reynolds stress model. This squared sectioned duct with roughened wall has been applied for coolant passage employed in gas turbine blades and heat exchanger in many engineering fields. Roughened wall is composed of many ribs, which is periodically located along the flow on the bottom wall of a square duct. In the numerical calculation, periodic boundary condition has been used for this flow and an algebraic Reynolds stress model is adopted in order to predict preciously Reynolds stresses. Calculated results of mean velocity and Reynolds stresses are compared with the experimental data in order to examine the validity of the presented numerical method and an algebraic Reynolds stress model. It has been pointed out as a characteristic features from the experimental result that secondary flow of the second kind is generated near roughened wall with flowing to roughened wall. The present method could predict its phenomenon correctly and reproduce well mean velocity profile distorted by the secondary flow. Besides, calculated results of Reynolds stresses are in good agreement with the experimental results, especially, shear stress distributions, which are characterized by generating the opposite sign regions, are predicted well as a result of estimating the secondary flow pattern preciously. Since the presented turbulent model is defined as high Reynolds number model, wall function is used for setting boundary conditions of turbulent energy and dissipation. As for this wall function, calculated results suggest that the universal constant of log-law velocity which is assumed in wall function, has to be changed according to the wall condition, that is smooth or rough.

View full abstract

A detailed measurement of the lateral vortex structure embedded in a wall boundary layer is presented to simulate the development process of the leakage vortex in the turbomachinery. The vortex was generated by a wall jet issuing perpendicularly to the mainstream from a slit located at the corner of the wall between two parallel ducts. The flow field was measured experimentally by a 5-hole probe utilizing a three-dimensional traversing device. Streamlines were traced to determine the leakage flow and through flow boundary, which not only explains the distribution of total pressure loss, vorticity, and other properties well but also clarify the flow structure in the vortex core. An angular velocity of the fluid particle was computed via eigenvalue of the velocity gradient tensor to represent the flow structure of the leakage vortex. The characteristic of the flow system is that the streamwise vortex is accompanied by two shear layers in which loss is almost generated. The vortex core forms a set of Bernoulli surface and is found to generate little loss in the vortex.

View full abstract

There is a continuing trend to wards building even larger wind turbines with energy outputs of several MW, especially for offshore use. In the design process of large-scale wind turbines it is essential to predict the performance of the turbine blades with great accuracy. The prediction through wind tunnel experiments is very difficult because the Reynolds number of the wind turbine model is much smaller than the large-scale wind turbine. Field site predictions are difficult due to wind speed fluctuations. These problems can be avoided by using CFD for predicting the characteristics of wind turbine blades. However, CFD has a weakness in predicting the drag and modeling the stall. In this paper two types of original CFD codes developed in the authors' laboratory are investigated for the flow around a wind turbine blade. The CFD codes are the SIMPLE algorism and the compressible method. The suitability of CFD for the prediction of wind turbine blade performance is studied by validating the computational results with experimental data.

View full abstract

In order to study the effect of turbulence sub-grid scale (SGS) component on particle motion, a new model named Dynamic Random Walk (DRW) SGS Coupling Model based on Eulerian-Lagrangian approach was presented. The advantage of the new model is that the Gaussian statistical distribution and local isotropic properties of turbulence SGS fluctuation magnitude could be parameterized by Germano's (1991) dynamic procedure. Using the present model, Large Eddy Simulation was performed for downward channel flow at Reynolds numbers of 180 that was identical to the DNS done by Rouson & Eaton in 1997. Through comparing of statistical properties of particles diffusion with DNS, the capability and limitation of presented DRW model was verified. Moreover, it was found that turbulence SGS component was strongly associated with particles motion because preferred particles were affected by preferred length scale of the eddy structure around, and fluid SGS component is indispensable in calculating particles motion with LES even though the particle Stokes number is high.

View full abstract

Computational technique for turbulent flows through two-dimensional steam turbine cascades are presented. A flux-splitting procedure is developed taking into account the thermodynamic properties of steam. The computational results of steam turbine nozzles are compared with the results with ideal gas properties. The mass flow rates are different from those based on ideal gas properties. However, non-dimensional characteristic values, which are mass flow rate coefficients, flow angles and energy loss coefficients, agree well with each other, where the specific heat rate of ideal gas computations is 1.3 in the superheated steam region and is 1.135 in the wet steam region. And it is cleared that assuming ideal gases with 1.135 for the specific heat rate of wet steam is appropriate in order to design blades and estimate the limited load condition in the supersonic flow region.

View full abstract

The flow around a sphere placed at various heights above a plane boundary has been investigated. The diameter of the sphere was d=57 mm, the gap between the sphere and the plane was various from 0 to 30 mm. The turbulent boundary layer thickness where the sphere was placed was various from 26 to 120 mm. The Reynolds number based on d was 8.3 × 10⁴. The flow visualizations by the smoke wire method and the surface oil-flow pattern method were carried out. The surface pressure distributions on the sphere and the plane were measured and the drag and lift coefficients were determined by integration of the surface pressure of the sphere. It was found that the effects of the plane, the gap and the thichness for the characteristics of the flow around the sphere and the fluid force acting on the sphere.

View full abstract

Transition process of internal gravity waves was regulated by giving small disturbance in a strongly stably-stratified mixing layer with a stepwise temperature distribution. The thermal disturbance composed of three frequency components, 1.1, 2.1, 3.2 Hz, sufficing the three-waves interaction condition of the internal gravity waves spontaneously generated in the unexcited stratified mixing layer was given from a fine heating wire placed at the position of the maximum local temperature gradient in it. Through the regulation, uncertain occurrence of the waves was locked up. Energy density levels of temperature and vertical velocity fluctuations corresponding to the internal gravity waves increased downstream and down-gradient heat flux along the local temperature gradient increased at the same time. The natural internal gravity waves collapsed in the downstream region and counter-gradient heat flux was observed in the lower side of the mixing layer. The enhanced internal gravity waves had greater potential energy and their collapse produced much larger counter-gradient heat flux.

View full abstract

The sound characteristics of the radiated waves from the open end of the duct with cavities have been studied numerically and experimentally for various depths of the cavities in connection with the sound reduction for the wide ranged frequencies. The sound pressure level (SPL), time history of the sound pressure and acoustic energy flux vector, i.e., instantaneous intensity, etc., are computed by a finite difference method. The present study shows that the lager depth of the nearest cavity from the duct yields the best configuration for the reduction of radiated power in the wide ranged frequencies. The present result also shows that the instantaneous intensity makes clear the interaction of the duct and each cavity. It is confirmed that the sound wave at the open end of the resonating cavity and the one of the duct are nearly equal in magnitude but reverse in phase. The computed results are in good agreement with experimental data.

View full abstract

An Aeolian tone, which occurs from the square cylinders with rounded corners, widely changes its frequency and sound pressure level according with the geometry of their cross sections. The purpose of the present research is to clarify the causal relationship with the cross section and the Aeolian tone, and to determine the cause of its frequency and sound pressure level. The results show that the frequency is affected by the size of the wake. In addition, the flow pattern around the cylinder is divided into two types according to whether the separated airflow reattaches on the body or not. The sound pressure level becomes minimal in the intermediate case of these two flow types, because the vortex shedding in this case is more unsteady and disorganized in space-wise than other cases.

View full abstract

Cavitation has been observed in a hydraulic oil flow separated from a smooth cylindrical wall, using various methods such as a microscope mounted with either a digital still camera or a high-speed video one, laser beam transmission, photo-multiplier and an electric charge detector. At the incipient phase of cavitation, a cavity suddenly emerged at the separation point with its upstream tip attached on the wall. Differently from cavitation previously observed in such flows as separated from an edge or a projection, however, the present incipient cavity neither stayed nor made a stable bubble on the wall, which was smooth this time. While drifting downstream, the newly born cavity was substantially enlarged, underwent severe deformation and ultimately vanished from the tail in as short a time as one-tenth of milli-second. When being enlarged, the cavity emitted flashing light similar to the one formerly observed in the cavitation on the needle projection. Transient electric charge synchronized with the light emission was also detected from the oil in the downstream. Moreover, the streamline extending from the separation point was spontaneously visualized, indicating heat generation at the separation point where significant shear stress acts. All these results seem to support the "rip-off" hypothesis previously advocated by the present authors.

View full abstract

In order to investigate the behavior of highly erosive, dangerous cavitation cloud, we newly build a closed-type cavitation-tunnel in our SIT, and systematically highspeed-phtographically observe such a behavior within the tunnel, for a somewhat wider water-temperature-range from 20°C to 45°C and from 13°C to 46°C, when the cavitation-shock-pressures have to be maximum with respect to the temperature, simultaneously measuring the cavitation nuclei distribution and Jacob number Ja, which have to affect the void fraction directly.

View full abstract

Among the recent measuring methods of multiphase flows, there are some methods in which we must mix fine particles into fluid as tracer or scattering particles ; for example, LDV, PIV and some methods using ultrasonic sound. The effect of fine particles mixed in on the flow characteristics is, however, not examined in detail. Therefore, in this study, we measured some parameters to investigate the effect of fine particles mixed into gas-liquid two-phase slug flows in a vertical pipe. Polyethylene particles of 10.6 and 160 μm mean diameter were used. The effects of particle mixing on average void fraction, large bubble and slug length, large bubble rising velocity and shape are discussed.

View full abstract

In this report, the effects of Reynolds number to the bubble-bubble interaction in a swarm of wall-sliding bubbles are investigated experimentally using PTV (Particle Tracking Velocimetry). Firstly, the drag coefficient of single wall-sliding bubble is measured for four kinds of liquid with different kinematic viscosity and is then compared to single bubble in an infinite liquid. Secondly, the difference of the bubble-bubble interaction among three types of Reynolds numbers, i.e. Re = 1.32, 2.83 and 15.42 is clarified using statistic data such as the interactive velocity vectors, the drag coefficient ratio and the presence frequency of the nearest bubbles. In particular, the discussion is focused on how the vorticity diffusion from the target bubble affects the nearest bubble.

View full abstract

For distilled water and mercury as working fluids, a drop formation was experimentally investigated at Bond number one order less than that of previously reported by others. A volume and a formative period of drop were measured for nozzles with outer diameters of 0.5 mm, 0.7 mm and 3.0 mm in altering nozzle exit pressure. To clear out a formative process of drop, high-speed video pictures were also taken to each nozzle for the distilled water. From above experiments, it was clear out that the drop volume at Bond number less thas 0.033 became almost constant without pressure dependence and was almost coincident with a critical volume estimated from a static equilibrium equation of gravitational pressure and surface tension. And it was also clear out that the reduction rate of the neck radius of the drop at near breakup was proportional to 1/(t_d-t) where t_d means the full time necessary to the drop formation and t means the elapsed time from the start of drop formation.

View full abstract

The effect of tip clearance on the transient process of rotating stall evolution in the axial compressor stage with three stator-rotor gaps was investigated experimentally by the pressure traces on the casing wall. The measurement of the pressure field on the casing wall covering the rotor and the velocity distribution behind the rotor with a slanted hot wire were performed near the stall point to clarify the effect of the rotor tip clearance on the flow field. In the case of small clearance, the stall evolution depends on the stator-rotor gap. However, in the case of large clearance, there is little difference of the stall evolution among three gaps. From the measurement of the pressure and velocity field near the stall point, it seems that breakdown of tip leakage vortex occurs in the case of large clearance for three different gaps. It is supported from the axial velocity distribution at the rotor exit. From the pressure field traces just before the stall, we can find the distinctive flow phenomenon like rotating instability which was found previously by Marz, et al. The tip vortex plays a very important role for stall inception.

View full abstract

A rapid expansion of moist air or steam in a supersonic nozzle gives rise to non-equilibrium condensation. If the latent heat released by condensation exceeds a certain quanity, the flow becomes unstable and a periodic flow oscillation occurs. In the present study, a numerical study of moist air flows in a supersonic nozzle was carried out using a special short duration supersomic wind tunnel, called a Ludwieg tube. The effect of the initial relative humidity of the moist air on static pressure, condensate mass fraction and nucleation rate and total pressure loss has been clarified in case without non-equilibrium condensation upstream of nozzle. The results obtained clearly show that the total pressure loss due to non-equilibrium condensation in the Ludwieg tube should not be neglected even for a very low initial relative humidity and it is resulted from the periodic excursions of the condensation shock wave.

View full abstract

The experimental study has been conducted on flow structure and heat (mass) transfer characteristics in the rectangular channels with a sharp 180-deg turn. Detailed mean and fluctuation velocity distribution were measured by using Laser Doppler Velocimeter at a Reynolds number, of 3.5×10⁴ and three dimensionless turn clearances (C) of 1.4, 1.0, 0.6. Based on present measured data with published local mass transfer results, the influences of turn clearance on the mass transfer and flow characteristics after the turn section are clarified. The characteristics of the flow reattachment and the separation bubble inside/after turn are much influenced by the size of the turn clearance.

View full abstract

We study a network model of crystalline membranes using Langevin simulation technique, where the model is defined by the Helfrich energy function. The Langevin equation is discretized by Euler algorithm with a discrete time step Δt. Time series of the bending energy and those of the mean square size are generated by iterations of the discrete Langevin equation. Comparing the time averages of these samples with those obtained by the canonical Monte Carlo, we see that the more Δt is small the more the Langevin results are close to the MC ones. Then by using several Δt's, the critical exponent of the shape fluctuation phase transition and the critical Hausdorff dimension are calculated in each Δt. As a result, we find that the critical exponent and the Hausdorff dimension obtained by the Langevin simulation are identical with those obtained by MC within the errors at Δt = 0.0001∼0.002.

View full abstract

Turbomolecular pumps (TMPs) are widely used in the semiconductor and thin film industries. In order to prevent corrosion on TMPs, it is sometimes coated with ceramic (SiO₂) film. The coated film changes surface roughness of the blades, and it is considered to influence on the pumping performance. The effect of surface roughness of the blades on the pumping performance is studied experimentally. First, the maximum compression ratio and pumping speed factor are measured for the non-coated TMP. Next, the TMP coated with SiO₂ is tested. The experimental results showed that the blade surface roughness changed by SiO₂ coating improved the pumping performance.

View full abstract

Effect of high frequency magnetic field on the natural convection in the Czochralski (CZ) silicon melt has been investigated numerically. The purpose of the study is to clarify the applicability of high frequency magnetic field for the control of the natural convection in the CZ silicon melt to produce high quality silicon single crystals. The numerically obtained results reveal that the melt natural convection is strongly affected by the high frequency magnetic field if the induction coil is placed at the slightly above the melt free surface. Although the convection structure is depend on the applied induction coil current and frequency, the high frequency magnetic field induces the opposite direction convection below the crystal compared with the natural convection if the adequate induction coil current and frequency is applied on the induction coil. The temperature field below the crystal is flattened due to the electromagnetically induced convection.

View full abstract

Direct numerical simulation (DNS) was performed for a non-isothermal air jet of the Reynolds number equal to 1 200 in order to reveal coherent structures of the jet. Fourth-order central finite difference was applied to the simulation. Effort was also made for experimental visualization (dye mixing and PTV) supporting the validity of instantaneous structures by DNS. Computational results for two kind of inlet profiles suggested that nozzle conditions scarcely affect turbulent statistics and coherent structures in jet-established-stage. Two-point correlations of velocity and temperature show similar distributions denoting that temperature can be used as a indicator of a vortex. Conceptual model of hairpin-shape vortex was proposed and validated by the two-point correlations and the PDF analysis for vorticity alignment; the hairpin-shape vortex stands with legs inclined downstream, and the inclination angle and the tilting angle between two legs are -45° and 40°, respectively.

View full abstract

Large eddy simulation (LES) was performed for a spatially developing round jet and its scalar transport at four steps of Reynolds number set between 1 200 and 1 000 000. Simulated domain, which extends 30 times nozzle diameter, includes initial, transitional and established stage of jet. Modified convection outflow condition was proposed in order to minimize effect of downstream boundary. Tested were two kinds of subgrid scale (SGS) models, Smagorinsky model (SM) and dynamic Smagorinsky model (DSM). In the former model parameters are kept at empirically deduced constants, while in the latter they are calculated using different levels of space filtering. Data analysis based on decay law of jet clearly presented performance of SGS models. Simulated results by SM and DSM compared favorably with existing measurements of jet and its scalar transport and however quantitative accuracy of DSM was better than SM at transitional stage of flow field. Computed parameters by DSM, coefficient for SGS stresses, C_R, and SGS eddy diffusivity ratio, Г_SGS, were not far from empirical constants of SM. Optimization of model coefficient was suggested in DSM so that coefficient C_R was nearly equal in established stage of jet but it reduced in low turbulence close to the jet nozzle.

View full abstract

The paper describes a three dimensional and time resolved temperature measurement technique for a gaseous flow and its application to a heating jet. For the whole field investigation, computed tomography is one of the significant techniques to make an internal structure clear and to understand the distribution of physical properties of flow fields. Conventional CT method, however, has difficulty in measuring the unsteady flow or turbulence due to the time consuming rotation of the projecting equipment. The objective of the present study is to develop the optical tomography technique with quad Twyman-Green interferometer in order to obtain the multiple finite fringe images from different directions simultaneously. In spite of the smaller number of projections, three dimensional temperature distribution of a heating gaseous jet from a rectangle nozzle was successfully reconstructed by the iterative calculation technique with a physical constraint for temperature. The technique was compared with the reliable thermocouple method on the steady jet and both measured results agreed quite well. The technique was further applied to the measurement of a periodic forced jet and obtained the instantaneous temperature field and their temporal fluctuation successfully.

View full abstract

This paper is the second report of the experimental study concerning the heat transfer in the separated flow behind a circular cylinder for Reynolds numbers from 120 to 30 000. In this report, unsteady and three-dimensional characteristics are described. Instantaneous temperature distribution and its fluctuating pattern on the cylinder surface were measured by an infrared camera. It was found that the pair of streamwise vortices formed in the near wake enhances the heat transfer around the rear stagnation of the cylinder. In particular, the streamwise vortices of 'mode-A', which is formed around Re = 200, effectively enhances the heat transfer around the rear stagnation. For 3 000<Re<15 000, the heat transfer around the rear stagnation increases remarkably with the increase in the Reynolds number. This is caused by the inception of the alternative reattaching flow on the rear face of the cylinder, interlocked by the vortex shedding behind the cylinder.

View full abstract

In order to reduce phase change time in latent heat technology, improvement of effective thermal conductivity of heat storage unit would be one of the techniques. Effect of effective thermal conductivity on melting time is studied analytically of circular composite heat storage capsules made by immersing phase change materials (PCM) into porous metals. Numerical and approximate analysis were made with the consideration of uniform and non-uniform heat transfer coefficients around the cylindrical surface. Four PCMs (H₂O, Octadecane, Li₂CO₃, NaCl) and three metals (copper, aluminum and carbon steel) were selected as specific materials. Porosities of the metals were restricted larger than 0.9 in order to lessen decrease in latent heat. Results show that reduction in melting time was obtained for the above PCMs, especially for low conductivity PCMs, Melting time obtained by approximate analysis agrees well with numerical analysis. High Nusselt number and high thermal conductivity of heat transfer fluid would be more effective to reduce phase change time.

View full abstract

Critical heat flux (CHF) has been measured on a small rectangular heated surface of L=5 and 10mm in length and W=4 and 6mm in width with an impinging jet of highly subcooled water to achieve an ultra high critical heat flux. The experiments were carried out at jet velocities of u = 5-60 m/s, system pressures of P=0.1-l MPa and degree of subcooling of ΔT_sub = 80-170 K. The effect of heater thickness appeared below 0.3mm and the CHF reduces with decrease in the thickness. Characteristics of the CHF for the longer heated surface L = 10mm agreed well with an existing correlation within ±20%. The maximum CHF of 212 MW/m² was recorded at P=0.5 MPa, u = 35 m/s and ΔT_sub=151 K. Then, the maximum CHF at atmospheric pressure can approach to 48% of the ultimate maximum heat flux q_{max, max} proposed by Gambill and Lienhard who evaluated it from the mean velocity of molecular dynamics.

View full abstract

A method for predicting evaporation performance of a new plate-type heat exchanger for water-refrigerant systems such as chillers has been developed. The main part of this heat exchanger consists of plates packed together in a casing, and winding tubes are connected to both sides of the plates. Refrigerant flows inside the tubes, and water flows in the space between the plates. A herring-born-like pattern is formed in this space by the cross sections of the winding tubes. This method estimates evaporation performance of the heat exchanger by using two new heat-transfer correlations : one for the winding-tube banks on the water side and one for pressure drops on both the water and refrigerant sides.

View full abstract

The effects of resonator on combustion oscillation stabilization have been investigated for gas turbine combustor, and mechanism of Helmholtz oscillation control by resonator has been discussed. Numerical model describing Helmholtz oscillation behavior, excited between combustor and fuel injector, is presented. And relaxation of oscillation amplitude corresponding to resonator parameters, such as throat diameter and length, are calculated and compared by atmospheric combustion experiments. In case resonator is resonant to combustor, it is shown that oscillation amplitude can be minimized. However, effectiveness of resonator is lessened with decrease of throat diameter or increase of throat length, while throat diameter and length are both tuning parameters of resonator. Throat resistance, which is determined by both of throat diameter and length, is actually one of the major parameters in Helmholtz oscillation control. Resonator parameters must be optimized in terms of attenuation characteristics as well as resonant frequency.

View full abstract

In order to improve the thermal efficiency of a micro gas turbine without increasing the temperature, a cycle which includes an isothermal expansion combustion process is proposed. The combustion should prefer entirely to be excuted in a turbine rotor and rather than in a stator to keep the total temperature constant. The cycle analysis indicates that, compared with the conventional gas-turbine cycle, the thermal efficiency of a constant total -temperature expansion combustion gas-turbine cycle is high because a high temperature field can be used for converting thermal energy to mechanical work. Furthermore, another cycle which includes a polytropic expansion combustion process and no isobaric combustion process has a high efficiency close to that of a constant total-temperature expansion combustion cycle. In addition, techniques of fuel injection are briefly discussed.

View full abstract

In order to examine the ignition process, direct numerical simulations have been carried out on spontaneous ignition of hydrogen gas jets injected parallel to a steady hot air stream. The simulations are based on a modified HSMAC method. Full chemical kinetics was used in the calculations. All the calculations were conducted under atmospheric pressure, in air flow velocity of 4 m/s and temperature of 1 000 K. The spouting velocities of fuel jets are 50 m/s and 10 m/s, respectively. Calculated results revealed that the ignition point moves from the peripheral region in the upstream of a fuel jet to its top region with a decrease in the jet velocity. The ignition takes place in the fuellean layer, where mixture fraction is 0.01 and scalar dissipation rate is suitable for the ignition.

View full abstract

The combination of a premixed flame from the main burner and a diffusion flame from the pilot burner is often used for industrial low NO_x burners. However, an oscillatory combustion can occur rather easily and is a big problem with this method. To investigate the problem, the correlations between combustion conditions and pressure fluctuation levels are studied on a swirling flow type combustor burning natural gas. It is found from the observation that an oscillatory combustion tends to occur for the case of having the wide region of higher OH concentration.

View full abstract

Homogeneous charge compression ignition (HCCI) combustion is regarded as next generation combustion method in terms of high thermal efficiency and low emissions. Natural gas is regarded as the most promising alternative fuel due to its clean burning and abundant supply. In this study, it was inventigated characteristics of autoignition and combustion of natural gas in an HCCI engine by using chemical reactions calculation and experimental study. And also, the operation conditions to realize combustion completion in an HCCI engine were suggested. The influences of n-butane blend ratio on autioignition and combustion was clarified in methane/n-butane/air mixtures. As a results of experiment and chemical reactions calculation, it becomes clear the following facts; i) Autoignition temperature of natural gas is in a range of 1 000±100 K without relations of equivalence ratios, intake temperatures and intake pressures, ii) To realize high thermal efficiency and low CO emissions, it is necessary to prepare operation conditions that maximum cycle temperature is over 1 500 K. iii) As blend ratios of n-butane are increased in Methane/n-butane/air mixtures, autoignition temperature and autoignition pressure become lower.

View full abstract

Exhaust emission and fuel consumption of a vehicle engine can be reduced by improving fuel atomization of an injection nozzle. Authors have already reported development of a new type of nozzle called Silicon Plate Nozzle (SPN) in order to improve fuel atomization by adopting micro-machining. This paper discusses combination of two SPNs which resulted in drastic improvement of fuel atomization. By combining two SPNs so that each rectangular aperture crossed with other, the fan-like spray was obtained. As a result, obtained Sauter Mean Diameter (SMD) of the fuel spray was 50 ?m or less at fuel pressure of 300 kPa. This experimental results were confirmed by numerical analysis.

View full abstract

Instantaneous and simultaneous measurements of two-dimensional temperature and OH-LIF profiles by combining Rayleigh scattering with laser induced fluorescence (LIF) are demonstrated in nitrogen-diluted hydrogen (H₂ 30% + N₂ 70%) diffusion flame interacted with a large scale vortex induced by the acoustic excitation at the fuel side. The dynamic behavior of the diffusion flame extinction and interruption during the flame-vortex interaction processes is investigated. The results obtained are described as follows. (1) When a large velocity fluctuation is given to the fuel jet, the temperature and the OH-LIF at the reaction zone become significantly lowered at the convex and circumferential part of the vortex where the thinning of the OH-LIF and the temperature layer and the consequent flame extinction are induced. (2) The burnt gas is pulled up from the upstream reaction zone by curling up motion of the vortex and high temperature region is formed inside the circumferential reaction zone. The high temperature gas dissipates with time and makes thick burnt gas region. The dissipated burnt gas reduces the supply of fuel into circumferential reaction zone and interruption of the reaction zone is induced.

View full abstract