An experimental study was carried out on the Reynolds stress distributions and production terms involved in the vorticity transport equation in a three-dimensional turbulence field of a stronger wall jet managed by an embedded streamwise vortex. Even in the strong three-dimensional flow field the mean velocity gradients 〓U / 〓y and 〓U / 〓z are of primary importance to the quantitative feature of the Reynolds shear stresses -(uv)^^^- and -(uw)^^^-, such as the maximal and minimal in their cross-stream distribution at a local streamwise position, respectively. A hypothesis of spanwise wandering of the introduced streamwise vortex was presented, to understand the large quantitative discrepancy in the stress-strain relationship between the Reynolds stress -(uw)^^^- and the rate of strain S_yz. The non-isotropic and non-uniform distributions of Reynolds stresses play a dominant role on the decay of the streamwise vortex. In a comparison with some results obtained in turbulent boundary layers disturbed by streamwise vortices, structural differences between the wall jet and the boundary layer reveal in the non-isotropy of the Reynolds shear stresses, -(uv)^^^- and -(uw)^^^-.

View full abstract

The rapid increase of available computational strage and speed has enabled the direct numerical simulation of the gas-particle jet, the Reynolds number of which is about one thousand, based on the Navier-Stokes equation and the Lagrangian equation of particle motion. In this study three-dimensional Eulerian air velocities and Largrangian particle trajectories are directly simulated to describe the effect of particle existence on the low Re number gas-particle free-jet flow using a two-way method. The results show that the small disturbances due to particle existence in the low Re number gas-particle free jet grow, and develop the flow to the turbulence. This is in fairly good agreement with the results obtained from the experimental flow patterns of low Re number free jets. The calculated flow characteristics of air and particles (mean velocity distributions and fluctuating intensity distributions) are also in fairly good agreement with experimental data obtained using a laser Doppler velocimeter. This means that the existence of particles enhances and develops unstable states in unstable laminar flows, such as low Re number free jets, to turbulence. The turbulence growth mechanism and the effects that the appearance of many small-scale eddies due to the particle existence has on the large-scales and vice versa have been discussed.

View full abstract

We have developed a new electrode-wire CT scanner for void distribution measurements. This CT scanner is an extension of the electrical probe method. The CT scanning method, which utilizes conductance values as projection data to use in the reconstruction of void distribution, is composed of electrode-wires stretched over a cross section of the flow area. The scanner is simpler and safer than conventional X-ray CT scanners. In this paper we describe the new CT scanner and discuss its applicability to void distribution measurements. Some examples of void measurements under both steady and unsteady two-phase flow conditions are shown.

View full abstract

In the present study an LDA-based technique was applied to analyze details of a transient pipe-flow induced upstream by a fuel ingector. Due to highly accurate measurements of the instantaneous time series of velocity, fuel flow rate and pressure gradient were achieved. The pipe flow has been driven under an injection pressure of 0.6MPa at high injection frequencies of 25 and 50 Hz. The influence of the outflow orifice diameter, varid in range from 100 to 500 μm, on the transient pipe flow dynamics has also been described. The results show that transient behavior of rapid oscillating pipe flow is remarkably and systematically dependent on injection frequency and duration as well as on the orifice diameter.

View full abstract

A high-pressure swirl type spray has been experimentally analyzed by using a phase Doppler anemometer(PDA). The experiment was performed with an injection pressure of 5 MPa and an injection frequency of about 46 Hz. The PDA data was obtained in a two-dimensional plane and arranged with the injection phase angle. The arranged data of mean velocity, Sauter mean diameter, and droplet number density were visualized by using commercial software. In this paper, the data were compared with the results of visualization. It is indicated by the experiment that the large velocity and large size droplets initiated from the injector with small spray angle at the beginning of the valve opening duration. This stage was called core jet. In the next stage called quasi umbrella spray, the spray developed with large spray angle. After closing the valve, the spray remains increasing its velocity and diameter. This was caused by the "post injection". Using PDA, the detailed feature of the spray can be discussed with high temporal resolution.

View full abstract

Time-dependent centerline velocity is measured using LDA-flow meter. Swirl injector at frequencies of 22 and 46 Hz generates oscillated flow under high-pressure from 4 up to 7 MPa. The encoded time-series of velocity is used to reconstruct the time-variable series of instantaneous flow rate and pressure gradient. Transient injection dynamics are analyzed to obtain basec characteristics of an injection system that will be able to be used for real time monitoring or active controlling of high-pressure gasoline injectors. The results obtained through analyzing the integrated mass rates within a cycle, indicate that a certain injection system is characterized by the flow rate constant γ which independent on injection duration and can be obtained from only accurate measurement of centerline velocity in the fuel transport pipeline. The present technique allows us to simplify the injection model and can be effectively adopted for practical employment on active and passive control with respect to gasoline direct injection engines.

View full abstract

Based on a quantitative analogy between laminar flows in curved pipes and orthogonally rotating pipes, a corresponding analogy of forced convective heat transfer is described for a region where both velocity and temperature fields are fully developed. Two types of thermal boundary conditions are studied:an axially uniform heat flux with a peripherally uniform wall temperature and an axially and peripherally uniform wall temperature. When the curvature parameter of curved pipe flow and the Rossby number of rotating pipe flow are sufficiently large, it is shown that the Nusselt numbers for these two flows agree with each other and the temperature profiles show strong similarity when the governing parameters and the Pradtl numbers of the two flows are equal. The remarkable effect of Prandtl number on heat transfer structure is demonstrated. Based on the computed results, semi-empirical formulae common to the two flows are presented for the mean Nusselt number.

View full abstract

The evaluation of convective heat transfer from fins to the air has been carried out in the case of annular fins subjected to an air flow parallel to the fin surface. The fin cooling is studied using infrared thermography. The thermal balance in a fin during its cooling process allows us to obtain the expression of the heat transfer coefficient from the temperature time evolution of the fin. Besides, the Particle Image Velocimetry allows us to obain the flow field in the mid-plane between two fins. The influence of the air flow velocity and of the fin spacing on the convective exchanges is studied. The tests have been carried out for Reynolds numbers(based on the shaft diameter and the air flow velocity)between 2550 and 42000, for different fin spacings. All the tests are correlated by an equation, expressing the mean Nusselt number on the fin as a function of the dimensionless fin spacing and of the Reynolds number.

View full abstract

Supersonic combustion using catalytic wire at constant temprature in a cold supersonic flow field was investigated in a square duct with a backward-facing step. The free stream Mach number was M_m=1.18. Hydrogen was injected transversely behind a backward-facing step into a cold air free-stream. The heat released from the catalytic combustion had no effect on the temperature of the catalyst. This indicates that the reaction rate of the catalytic combustion observed in this study was determined by the concentration of H₂ and / or O₂ on the surface of the catalyst. The spatial distribution of heat released from the catalytic combustion in supersonic turbulent mixing layer, corresponds to the spatial distribution of concentration of H₂ and / or O₂ in local, was obtained. It was found that the most suitable position for supersonic combustion was at the outer edge of the mixing layer.

View full abstract

An numerical simulation of hydrogen-oxygen reaction over platinum catalytic surface was carried out and the effect of surface diffusion velocity on catalytic reaction was studied. In this simulation, no convection is assumed to simplify the analysis. The numerical model is as follows;A spherical platinum of 0.75mm in radius is surrounded with the mixture. Three based partial differential equations, which govern mass, energy and species concentrations, are stated in one-dimensional polar coordinates. Axial simmetry, zero azimuthal velocity, and no gradient along the axise ar assumed. Because of low Mach numbers, constant pressure is assumed, and therefore, momentum equation is not used. The Langmuir-Hinshelwood mechanism was used for the catalytic reaction model. The adsorbed species are H, O, OH and H₂O and 3 elementary surface reactions are used. As the results, a steady state is observed. The surface temperatures at the steady state do not depend on surface diffusion velocities but on adsorption rates and desorption rates. The catalytic ignition temperatures increase as H(s) surface diffusion velocity decrease. They are independent on surface diffusion velocities of O(s) and OH(s).

View full abstract

This work summarizes numerical results for a diffusion flame formed from a cylindrical tube fuel injector, issuing a gaseous methane jet vertically in a quiescent atmosphere. The primary objective is to predict the flame base height and other flame characteristics as a function of the fuel jet velocity. A finite volume scheme is used to discretize the time-averaged Navier-Stokes equations for the reacting flow resulting from the turbulent fuel jet motion. The turbulent stresses, and heat and mass fluxes are computed from a Reynolds stress turbulence model. A chemical kinetics model involving two-step chemistry is employed for the oxidation of methane. The reaction rate is determined from a procedure which computes at each point the minimum(process limiting)rate from an Arrhenius(kinetically controlled) expression and the eddy dissipation(turbulent mixing controlled)model. The Reynolds stress model(RSM), in conjunction with the two-step kinetics and the eddy dissipation model, produces flame base height and other flame characteristics that are in good agreement with experimental results. Numerical results are also in agreement with the hypothesis of Van Quickenbourn and van Tiggelen concerning the stabilization mechanism of lifted diffusion flames. The present results show the existence of the condition of tangency, as postulated by Van Quickenbourn and Van Tiggelen, between the jet axial velocity and the flame velocity profiles at the flame base. Furthermore, the burnout rate calculations indicate a high degree of premixing of air and fuel upstream of flame base for moderate to high fuel jet velocity cases.

View full abstract

In this study, a new ceramic-opposed jet burner was designed to ensure the adiabaticity of the flame. In the opposed-jet-turbulent premixed flame, the concentrations of various stable species as well as the mean flame temperature distributions and structures have been investigated. In the present burner, turbulent premixed flame can persist far beyond the stretch rate at which the wrinkled laminar flame is usually extinguished. The results show that the flame structure consists of the stretched stringy vortices involving variou sreacting species(e.g., C₃H₈, hydrocarbons, O₂)which are referred to as the distributed reaction zones, and that the flames are divided into two categories:(1)transition flame from a wrinkled laminar flame to a distributed reaction zone, and(2)a fully developed distributed reaction zone. The structure of the distributed reaction zone depends mainly on the mixture jet velocity and is independent of the equivalence ratios. The stretched stringy vortice of the distributed reaction zone eradicates the nonuniformity of temperature and concentration fields. The hydrocarbon species as well as fuel are completely consumed in the distributed reaction zone, and almost complete combustion is achieved.

View full abstract

A numerical study is performed on a two-dimensional, chemically reacting boundary layer to achieve a physical understanding of the stabilization and triple structure of a bulk flame. A theoretical model is presented to couple fluid dynamics, chemical kinetics and heat transfer at the surface of a flat wall. Computations are carried out at different Nusselt numbers and Reynolds numbers based on the free stream velocity. When the Nusselt number is increased, the flame apex begins to leave the wall surface and this manifests an unstable process. Finally it is blown off when the Nusselt number becomes sufficiently large. For a stable flame, three zones can be identified near the flame apex:a nonreactive mixing zone, a premixed / partially premixed flame zone and a diffusion flame zone. When the Reynolds number is decreased, the reaction zone develops away from the wall and more heat and fuel mass are transferred upstream of the burner, which is conducible to the stabilization of a bulk flame

View full abstract

For an application of DFWM to temperature measurement of gaseous flows, we detect the DFWM spectra of iodine vapor seeded in argon and investigate fundamental properties of the spectra. The DFWM signal can be detected with the high S / N by the use of different polarizations for the input laser beams. Shapes of the DFWM spectra depend on the input laser intensity and pressure, mainly due to the change of ratios of input laser intensity to the saturation intensity. The spectra also depend on temperature, because the population of iodine vapor varies with temperature, showing the feasibility of the temperature measurement of the gaseous flows.

View full abstract

Planar laser-induced fluorescence(PLIF) using OH and NO facilitates noncontact two-dimensional temperature measurements, and is therefore expected to be an effective technique for combustion thermometry. However, the accuracy of the measurement has not been clarified. The relaxation in the upper rotational levels of the tracer molecule is not easy to be predicted, and the rotational dependence of the molecular constants, such as absorption and emission coefficient, has not been fully clarified. Furthermore, it is not confirmed whether the rotational distribution of the tracer molecule maintains the quasi-equilibrium states or not. In this study, we investigated the accuracy of thermometry by means of laser-induced fluorescence (LIF) using NO:A-X(0, 0)and OH:A-X(3, 0) excitation for a premixed methaneair laminar flame in atmospheric pressure as a measuring target. These were checked by narrow-band N₂ CARS with high spectral resolution, which is believed to be the most reliable temperature standard in flames. NO LIF temperature showed several % lower temperature than N₂ CARS one, and OH LIF temperature 30% lower. Both methods were calibrated under exponential dependence assumption of molecular constans, and the calibrated temperature of NO and OH LIF agreed well with N₂ CARS one. Furthermore, two-dimensional thermometry in a flame cross section using calibrated NO PLIF was demonstrated.

View full abstract

In this study, firstly we endeavor to prescribe logic definitions and reasoning process for obtaining correct rocket power and relevant efficiency equations. With the Lagrangian Reynolds transport approach, we also rigorously derive these highly generalized equations for rocket total kinetic power, thrust power and related propulsive, thermal, and overall efficiencies. They involve a few more physical effects including rocket acceleration, relative flow velocity / steadiness, outlet pressure, and gravity, thus open an original route for improving rocket propulsion analysis / design. Incidentally, it is interesting to note that under some conditions, a rocket in flight may retain less kinetic energy due to ejecting more kinetic energy of burning propellant per second than the thrust power acquired. Also, the derived correct rocket total kinetic power equations are quite reasonable in that all the velocities involved are of relative nature, thus invariant with respect to different observers and shielding the possibility of violating energy conservation law.

View full abstract