Search

Distributions of velocity and concentration of turbulent gas jets with various densities from helium to carbon dioxide against ambient air are measured by using particle image velocimetry and laser induced fluorescence methods. The turbulent transport of the jet fluid is directly investigated to analyzing the distributions of jet-fluid concentration and its vortex motion. The result shows that the self-similarity exists between profiles of scalar flux and Reynolds stress along the radial direction. Furthermore, several subgrid scale scalar-flux models for large eddy simulation are discussed. Based on the scattering plots of model values of subgrid scale scalar-flux against the measured values, the characteristics of each model are basically similar for every density conditions. Also, the Favre-averaged scale similarity model performs well for wide various densities.

The flow of a high-speed unsteady jet is analyzed using computational fluid dynamics for an incompressible flow with the k–e turbulence model. The pseudo-nozzle concept is applied to the inlet condition with a large pressure gradient. The results show that the time history of the jet development agrees with the experimental data for methane and hydrogen fuels. In addition, the effect of the injection condition on the development of the jet tip is well described with this model. Furthermore, the effects of inlet conditions of the turbulence intensity and scale on the flow are investigated. © 2006 Wiley Periodicals, Inc. Heat Trans Asian Res, 36(1): 1– 12, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/htj.20141

The flow and mixing process of unsteady jets are fundamentally analyzed by using large eddy simulation (LES). The effects of nozzle velocity and turbulence intensity on the turbulent eddy structure and mixing process between nozzle fluid and ambient fluid were investigated. The results show that toroidal-shape vortex, which is growing up around jet tip, mainly accelerate the entraining flow. Also, increasing the turbulence intensity in the nozzle encourages the mixing in the jet without change of the jet-contour. Furthermore, when the rise-up time of initial nozzle velocity is elongated, turbulent mixing is suppressed.

Effect of hydrogen addition to hydrocarbon fuels on combustion characteristics were investigated using chemical full kinetics. The changes of ignition delay time and laminar flame velocity when hydrogen is added to methane or propane fuels were calculated. The result shows that for ambient pressure condition, ignition delay time decreases corresponding to the amount of hydrogen addition. However for high pressure condition, when 10% hydrogen is added to hydrocarbon fuel, ignition delay time decreases as short as that of hydrogen itself. Laminar flame velocity increases gradually with adding hydrogen, and when more than 70% hydrogen is added, laminar flame velocity rapidly increases. Furthermore, the production rate of OH is an important factor for the change of ignition delay time and laminar flame velocity, when hydrogen is added.

The combustion process of natural gas in a premixed charged compression ignition (PCCI) engine is analyzed using computational fluid dynamics included in heterogeneity of mixture concentration and temperature with stochastic approach. The non-uniform states of turbulence mixing, ignition process is statistically described using the probability density functions (PDF). The results show that the course of in-cylinder pressure is good agreement with experimental data and the effect of mixture heterogeneity on the ignition delay and the rate of heat release is revealed.

The velocity field and the temperature distribution are measured simultaneously in a nonpremixed jet flame by applying particle image velocimetry (PIV) and planar laser induced fluorescence (PLIF) of nitric oxide with known concentration added as a tracer. The fuel flow of low temperature is surrounded by the high temperature region. Also, turbulence kinetics in a jet flame was analyzed based on the statistical characteristics of the temperature and velocity, including mean velocity, turbulence intensity and Reynolds stress. It was shown that u' is quickly decayed in the high temperature region. And the maximum point of density weighted average of Reynolds stress is located at fuel stream side of shear layer. As a result, production of the turbulence is mainly occurred in the fuel stream.

To evaluate the promotion of turbulent mixing, the time-average and fluctuation of fuel concentration were measured in a mixing channel with triangular plates using Rayleigh scattering method. Based on the obtained results, the deviations of fuel distribution were calculated at some crosssections, and also probability density functions were estimated by comparing with measured deviation and predicted ones from turbulence-mixing model. Furthermore, the effect of mixture heterogeneity on NO concentration in a burnt gas was demonstrated, and the trade-off relation between a mixing evolution and an increase in pressure drop was exhibited.

Cross-correlation PIV is repetitively applied to obtain the time-change of instantaneous velocity vectors, which enables us to analyze turbulence kinetics. Each term in the balance equation of turbulence energy is evaluated, based on the ensembles of fluctuating velocities and instantaneous velocity-gradients, thereby finding primary factors of production and dissipation of turbulence energy. It is shown that turbulence production rate is high in the shear layer because of a large Reynolds stress, and viscous dissipation of turbulence energy occurs mainly near the jet axis. Furthermore, typical turbulence eddies in the shear layer are sampled to discuss the effect of instantaneous vortex-motions and fluctuating flows on turbulence production. It is observed that the pair of counter-rotating vortices appears and accelerates the flow in between, resulting in an increase in quantities of instantaneous Reynolds stress at a particular direction.

Time changes of a two-dimensional distribution of velocities in a nitrogen jet and a methane jet flame are measured by cross-correlation particle-image-velocimetry (PIV). From the measured results in a jet and a jetting flame, it is shown that the velocity gradient at shear layer in the reacting zone is increased due to the local acceleration by buoyancy, resulting in higher turbulence intensities compared with those in a non-reacting jet. Also, from the change of the distribution of velocity vectors with time, it is observed that the turbulence eddies are carried downstream along the gas motion with a little transformation. The time scale of turbulence at every location in the flow is obtained from the autocorrelation function of the velocity fluctuations. Furthermore, this can afford to estimate the turbulence length-scale. It is shown that the characteristic length-scales of a flaming jet are about 1.5 times greater than those of a non-flaming jet.

A fully three dimensional numerical model for the combustion process in a spark-ignition engine was established using the combustion submodel described of the flame area evolution. Computations carried out the process of premixed combustion and they were compared with the results using the combustion submodel of the eddy break-up. The combustion in a disc chamber of constant volume was simulated. The results show that the course of flame propagation is well predicted. The calculations for different conditions in a natural-gas engine are performed. It is shown that the effects of fuel-air equivalence ratio and swirl intensity on flame propagation are described. The cylinder pressure histories are agreed with measured one, and the distributions of velocity vectors and temperature in a cylinder are revealed.

A computational fluid-dynamic simulation for a turbulent nonuniform combustion is established using a stochastic approach. At each point in a turbulent flow field, the variations of species mass fraction and temperature are statistically described by the joint-probability density function (pdf), and the velocity variation is expressed using the conventional k–ϵ turbulent model. The transport equation of this joint pdf of mass fraction and temperature is calculated by a finite-difference method in convection and turbulent diffusion and by the Curl collision-redispersion model in molecular mixing. This method is applied to simulate the process of scalar dispersion in a uniform isotropic turbulent flow. The results show that the profile of an averaged scalar is quite similar to those calculated using conventional transport equations. Furthermore, a reasonable degree of reproduction is achieved for the pdfs of the scalar at each point in the flow field. © 2001 Scripta Technica, Heat Trans Asian Res, 30(6): 503–511, 2001

Numerical simulations were performed to predict fluid motions induced by interactions between density and pressure gradients. The presence of a vortex string situated in a deviated position from the jet flame axis generates two pairs of vortices near the flame front where a significant density change exists. It is predicted that the deformation and stretching of the flame front take place once the vorticity production becomes stronger than the dissipation due to viscosity. Vorticity production is caused by the baroclinic effect in the upstream position, but the generation due to gas expansion supersedes that of the baroclinic effect in the downstream position. When the pressure gradient is greater than a certain critical value, the vorticity production proceeds at a higher rate, which promotes the deformation of the flame front and accelerates the heat release.

The flow instability of a jet diffusion flame was investigated based on a linearized stability theory. The equation of disturbance was solved numerically for a two-dimensional parallel flow with density variation. Results show that the existence of a hot layer at the jet boundary makes the flow more stable, increasing the critical wave number and critical Reynolds number at which turbulence neither grows nor declines. The increase of the critical wave number is due mainly to the interaction between the flow and density gradient, and the local change of the viscous dissipation increases the critical Reynolds number, thereby leading to a wider stability limit.