F. Robben

Translational nonequilibrium in free jet expansions

The velocity distribution function for helium in translational nonequilibrium in hypersonic free jet expansions has been investigated with an electron beam fluorescence technique. A unique computer controlled Fabry‐Perot interferometer and data acquisition system has been used to observe the 5015.67 A radiation from the 31 P‐21 S transition in helium. From the spectral structure, the Doppler broadening and shift of the line due to the velocity distribution function can be determined. With this technique parallel and perpendicular gas kinetic temperatures, mean flow velocity, and relative density measurements were made along the centerline of the free jet, from near equilibrium conditions through the midpoint of transition (T‖/T⊥ = 2.0). A comparison with kinetic theory predictions of the temperatures and velocity distribution functions indicate that the adoption of the ellipsoidal model, although sufficient for fitting the measured distribution function, is inadequate for determining the parallel temperat...

Temperature and density in a hydrogenair flame from Rayleigh scattering

Abstract A technique has been developed for obtaining spatially resolved measurements of temperature and density under combustion conditions. The spectral structure of Rayleigh scattered laser light is determined using a Fabry-Perot interferometer. Spectra in the post-combustion zone exhibit simple Doppler broadening with a Gaussian line profile; however, spectra in the higher density, precombustion zone have appreciable Brillouin components. Radial and axial surveys of temperature and intensity in the premixed hydrogenair flame issuing from a 3.2 mm diameter welding torch tip are obtained. Measured temperatures in the post-combustion zone are in agreement with the adiabatic flame temperature.

Rotational temperatures in nonequilibrium free jet expansion of nitrogen

Rotational nonequilibrium was investigated in hypersonic free jet expansions of nitrogen using the electron beam fluorescence technique. The results confirm the conclusion of previous investigations that a dipole excitation model, with an assumed Boltzmann energy distribution, is not consistent with the measured line intensities. This discrepancy was examined quantitatively and found to be independent of density and source Knudsen number (except at number densities greater than 1016 cm−3). The effect was attributed to interactions with the ejected (ionized) electron and a new excitation model was developed and shown to be consistent with the measurements throughout the flow conditions explored. The resulting rotational temperatures were compared to a simplified relaxation model of the jet and indicate a rotational collision number of 1.9.

Density fluctuations of flames in grid-induced turbulence

Abstract Density fluctuations have been measured using Raleigh scattering in the heat release zone of a rod-stabilized C 2 H 4 /air flame propagating into a grid-induced turbulent flow field. Probability density functions show that intermediate states between burned products and cold reactants are significant in determining density statistics. A comparison of the statistics for fluctuation intensity with results predicted by the model of Bray, Moss, and Libby, in which intermediate states are neglected, further confirms this point. However, the skewness is accurately predicted by the Bray-Moss-Libby model. Since intermediate states are approximately symmetrically distributed, they do not contribute significantly to the skewness.

Catalyzed combustion of H2/air mixtures in a flat-plate boundary layer: I. Experimental results

Abstract A study has been made of the combustion characteristics of lean hydrogen-air mixtures flowing over a heated catalytic platinum plate. The objectives of the investigation were to develop a better understanding of the interaction between fluid mechanics, gas-phase combustion, and surface reaction in high-temperature surface-catalyzed combustion. The experimental system consisted of a thin quartz plate with vacuum-deposited platinum heating strips mounted over an open atmospheric-pressure jet of premixed hydrogen and air. Boundary-layer density profiles were measured using differential interferometry for flow visualization studies and Rayleigh scattering for point-density measurements. Laser Doppler velocimetry was also used to measure the velocity distribution. The presence of heat release due to surface reaction was determined from changes in heating-strip power inputs with fuel addition. Results were obtained for a range of equivalence ratios from 0.05 to 0.3 and plate surface temperatures from 470 to 1300 K. Significant surface heat release was found for all mixtures at plate temperatures as low as 470 K. At increased equivalence ratios and plate temperatures heat release due to gas-phase reaction was present. This was characterized by local increases in temperature across the boundary layer and an increase in thermal boundary-layer thickness. Comparisons with numerical calculations based on a surface reaction model which included finite-rate surface oxidation of H2 and radical recombination indicated qualitative agreement in general boundary-layer behavior, but the predicted gas-phase heat-release rate was considerably higher than found experimentally.

Experimental study of combustion in a turbulent boundary layer

H 2 -air combustion in a heated turbulent boundary layer has been studied through time and space resolved measurements of velocity (laser Doppler anemometry) and density (Rayleigh scattering). The mean, rms fluctuation, probability density function and spectral density distribution of these measurements have been obtained. The experimental conditions covered equivalence ratios ranging from 0 to 0.2, wall temperatures from 1100 to 1300 K and free stream velocities from 14 to 22 m/sec. The nominal boundary layer Reynolds number based on the free stream conditions was 1·1×10 5 . Three modes of combustion in the boundary layer were found: 1) surface reaction, 2) surface and gas phase reaction in the boundary layer, and 3) a flame-like region situated at the edge of the boundary layer. Velocity fluctuations of about 6% in the non-heated boundary layer were increased to 10% by wall heating. When combustion occurred the turbulence intensity decreased, to a minimum of 2% at the highest equivalence ratio. The suppression of turbulence intensity by combustion appears to be associated with the decreased density, and lower effective Reynolds number, in the region of wall shear. Density fluctuations of 8 to 10%, both with and without combustion, were found in the region of maximum density gradient. Under conditions where a flame-like region occurred, it was concluded from the spectrum of the density that the region oscillated at a fairly regular rate ranging from 400 to 2000 Hz.

Two-point rayleigh scattering measurements in a V-shaped turbulent flame

A Rayleigh scattering technique for simultaneous measurements of density at two points in a flame was developed. This technique was used to deduce the spatial and time-space correlation and the two-point joint probability density functions of the density in a premixed turbulent flame in three orthogonal directions. A grid generated turbulent V-shaped flame at a flow velocity of 7.0 m/s and a half angle of 12 degree was used. The analysis show that this flame consists of structures of equal length scale in all three directions which decay as they are convected downstream by the flow. It is also shown that this flame does not consist of flamelets.

COMBUSTION-TURBULENCE INTERACTION IN THE TURBULENT BOUNDARY LAYER OVER A HOT SURFACE

The turbulence-combustion interaction in a reacting turbulent boundary layer over aheated flat plate was studied. Ethylene/air mixture with equivalence ratio of 0.35 was used. The free stream velocity was 10.5 m/s and the wall temperature was 1250° K. Visualization of combustion structures was provided by high-speed schlieren photographs. Fluid density statistics were deduced from Rayleigh scattering intensity measurements. A single-component laser Doppler velocimetry system was used to obtain mean and root-mean-square velocity distributions, the Reynolds stress, the streamwise and the cross-stream turbulent kinetic energy diffusion, and the production of turbulent kinetic energy by Reynolds stress. The combustion process was found to be dominated by large-scale turbulent structures of the boundary layer. Combustion caused expansion of the boundary layer. No overall self-similarity is observed in either the velocity or the density profiles. Velocity fluctuations were increased in part of the boundary layer and the Reynolds stress was reduced as compared with the isothermal case. The turbulent kinetic energy diffusion pattern was changed significantly, suggesting that a modification of the boundary layer assumptions will be required when dealing with this problem numerically.

Conditional velocity statistics in premixed CH4-air and C2H4-air turbulent flames

Conditioned fluctuation intensities and Reynolds stress have been determined separately for the reactants and for the products in two methane-air and two ethylene-air v-shaped premixed turbulent flames. The conditioned quantities are deduced from velocity data obtained by a two-color LDA system using a conditional analysis method based on the presence of a thin flame sheet dividing the two regions. Within the turbulent flame brush, the conditioned turbulent kinetic energy in the reactants remains at the free stream level. The conditioned turbulent kinetic energy in the products, k p , are higher than in the reactants indicating flame-generated turbulence. Also, k p increases with distance from the flame stabilizer, x . Nevertheless, the two conditioned Reynolds stresses are insignificant. The unconditioned turbulent kinetic energy and Reynolds stress increase with x . This increase is proportional to the increase in the difference between the conditioned mean velocities Δ U . The changes in Δ U suggest expansion stream tube across the flame sheet.

Density fluctuations in premixed turbulent flames

The simultaneous two-point density fluctuations in a V-shaped turbulent flame are measured using a two-point Rayleight scattering method. A wrinkled laminar flame model with finite instataneeous flame thickness is developed for the flames studied. The reaction front spatial probability density function is both measured directly and also calculated from the measured mean density. An analytical expression for this pdf is given which is derived based on a thin flame model. The mean, rms and correlation coefficients are calculated using the finite reaction front thickness model and the results are compared with the experimental data. The pdf of the intermediate states are shown to be due to the reaction front thickness.