Scalar Length Scales Research Articles

Premixed flame front structure in intense turbulence

Six methane/air flames of equivalence ratio 0.7, Ka =1–17 stabilized on a intense turbulence low-swirlburner of were investigated by Rayleigh laser sheet measurements of the gas density within the preheat zone (approximately 400–1000 K). Lean premixed turbulent flames with Karlovitz numbers Ka , greater than unity are becoming of increasing practical importance but only limited detailed experimental data are available. This regime is often called the distributed reaction zone and a new approach to analyzing flame front broadening was developed to address issues of flame three dimensionality. Instantaneous progress variable, c , contour sets constitute the basic data and a statistical analysis of three aspects of this data are presented: (1) c contour spacing (2) flame front curvature, and (3) flame surface density. A detailed investigation revealed little change of these parameters with turbulence, even at the highest Ka numbers: the contour spacings were similar to those derived from laminar flame calculations and neither the flame front curvature nor the flame surface density varied significantly with c . Geometric effects of turbulence on flame front structure still predominate at these turbulence levels. The primary effect of increasing turbulence is increasingly to convolute the flame front and so increase the flame surface density, reduce the scalar length scales, and increase the burning rate. This finding is also reflected in the significant broadening of the flame front curvature distributions with increasing u′/S L . The data presented here indicate that the diagrams often used to delimit the various regimes of premixed turbulent combustion and hence guide modeling are based on parameters which need to be carefully investigated to determine their appropriateness. The results are consistent with an interpretation of combustion/turbulence interactions which sees the reaction zone thickness as the significant scalar length scale.

The use of photoactivatable fluorophores in the study ofturbulent pipe mixing: effects of inlet geometry

The effects of inlet conditions on downstream mixing inturbulent pipe flow have been studied with the use of photoactivatable fluorophores and standard laser-induced fluorescencetechniques. The different inlet conditions included both geometrychanges and changes in the manner in which the constituents wereintroduced into the flow. Results indicate that small changes ininlet geometry can greatly affect the downstream mixing rate.Changes in the geometry of the inlet had a greater influence ondownstream mixing than did the manner in which constituents wereintroduced into the flow. Further experiments included a staticmixer that was used in conjunction with two different inletconditions. It was found that the inlet condition greatly affectsthe effectiveness of the static mixer. The static mixer is mosteffective when placed downstream of an inlet that produces scalarlength scales that are similar in size to the elements in the mixer(i.e., the pipe diameter). Overall, the results of theseexperiments demonstrate that the method used to introduce twoconstituents to be mixed in pipe flow can profoundly affect thedownstream mixing rate.

One-dimensional raman scattering for determination of multipoint joint scalar probability density functions in turbulent diffusion flames

Joint probability density functions (PDFs) are one approach to describe turbulent reacting flows. One-dimensional spontaneous Raman scattering (1D-SRS) allows the simultaneous detection of the scattered light from numerous chemical species in a reacting flow at different points along a line, thereby providing all necessary information for multipoint joint scalar PDFs. The flow analyzed in this work with the help of 1D-SRS is a turbulent methane diffusion flame with a coflow of oxygen at moderate turbulence intensities. The simultaneous measurements of different species concentrations at several points determine multipoint joint scalar PDFs. From these PDFs, statistical moments such as mean values and standard deviations can be derived, as well as scalar spatial correlations and scalar length scales. The mapping of turbulent structures in the reacting flow is possible with all the statistical information contained in multipoint joint scalar PDFs. The chemical reactions in the methane flame correspond to a global two-step mechanism. This can be seen from the map of the mean mole fractions and is confirmed by the map of species correlations. The scalar length scales calculated from the space-correlation functions of concentrations show the general course of length scales for velocities in the similarity regions of the jet. However, length scales will be different for different chemical species and they deviate from length scales for velocities in regions of the jet where high turbulent reaction rates prevail. Scalar spatial correlations and length scales are directly reflected by the multipoint scalar PDFs that may serve as a data basis for mixing models in PDF transport equations.

Flame surface density and burning rate in premixed turbulent flames

The flame surface density has been measured in hydrocarbon/air stagnation point and v-shaped premixed turbulent flames. A method is proposed to determine the flame surface density from the data obtained by laser sheet tomography. The average flame length and flame zone area as a function of the progress variable are calculated from a map of progress variable and a set of flame edges obtained from the tomographs. From these results a surface density estimate in two dimensions is determined. By this technique it is possible to avoid the difficulties which arise when using an algebraic model based on the measurement of the flame front geometry and a scalar length scale. From these results the burning rate can be obtained which compares well with estimates calculated using the fractal technique. The present method, however, is not constrained by a minimum window size as is the case for the fractal determinations.

The spatial scalar structure of premixed turbulent stagnation point flames

A high speed tomographic technique is used to quantify the spatial scalar structure of stagnation flow stabilized premixed flames. Studies are performed on CH{sub 4}/Air and C{sub 2}H{sub 4}/Air turbulent flames with equivalence ratios ranging from 0.75 to 1.0. The gas velocity at the nozzle exit is 5 m/s, the turbulence intensity is 7%, the integral length scale 3 mm and hence the turbulence Reynolds number is 70. The light source is a copper vapor laser which produces 20 ns, 5 mJ pulses at a 4 KHz repetition rate. Cylindrical lenses transform the 38 mm circular laser beam to a sheet 50 mm high and 0.6 mm thick. A high speed Fastax camera is used to record the tomographic images formed by the scattering of light from oil droplets seeded in the reactant flow which disappear at the flame front. The films are digitized and the wrinkled flame front extracted from the images by a thresholding technique. The integral scalar length scales proposed in the model of Bray-Champion-Libby are deduced from the tomographic movies by numerically constructing flame crossing sequences from the burned and unburned probability density functions of flame crossing length and performing a spectral analysis of these sequences.more » The integral scalar length scales are found to be similar to the integral length scales of the incident turbulence. A fractal analysis was performed on the flame boundaries to characterize the flame geometry and provide from an estimate of the flame surface area a measure of the turbulent burning rate. The flame area increase is found to underestimate the turbulent burning velocity when compared with direct measurements of the flow velocity at the cold boundary. 12 refs., 2 tabs.« less

Scalar mixing in isotropic turbulence

Eswaran and Pope [Phys. Fluids 31, 506 (1988)] performed direct numerical simulations to study the influence of the initial scalar integral length scale on mixing in stationary, isotropic turbulence. Their data demonstrate that both the decay rate and the shape of the rms versus time curve depend on the initial value of the scalar-to-velocity integral length-scale ratio. The present paper discusses modifications of the high Reynolds number theory of Corrsin [AIChE J. 10, 870 (1964)]. The predictions mirror the behavior found in the moderate Reynolds number simulations.

Simulations of scalar mixing in grid turbulence using an eddy-damped closure model

A physical-space eddy-damped quasinormal Markovian model for the two-point scalar correlation function is used to simulate wind tunnel experiments on scalar mixing. The simulations closely approximate experimentally observed trends of scalar variance and length scales in both heated-grid and heated-screen experiments. The simplicity of the model should enable its effective use in modeling the analogous turbulent mixing terms in the two-point scalar probability density function equations.

Scalar length scale variations in premixed turbulent flames

A general formulation and closure of a two-point two-time covariance balance equation is applied to the problem of calculating scalar length scale variations in premixed turbulent flames. Situations are considered in which the turbulent combustion zone consists either of a single thin wrinkled flame or of packets of reactants and products separated from each other by thin flames. The objective of the present work is to investigate the variation with location within the turbulent flame of characteristic dimensions of these packets or flame wrinkles. A normal or oblique planar flame configuration is considered. A similar formulation has previously been applied to calculation of time scales in premixed turbulent flames. Predicted packet passage time pdf's, autocovariances, mean packet time variations and power spectral densities all show satisfactory agreement with experimental data. However, the predicted time scales do not provide much information about the shape and size of the burning packets or wrinkles. The present paper includes predictions for both normal and transverse scales of these packets or wrinkles. It is found that the normal and transverse dimensions vary with position in different ways, resulting in an anisotropy of the scalar field. Insufficient experimental data has been found to permit a quantitative comparison with these predictions. However, where a qualitative comparison is possible, this lends support to the results of the analysis.