OCR Text |
Show Presented al: the Fifth Imerna/ionai Symposium on the Applications o/Laser Techniques to Fluid Mechanics, Lisbon, Portugal, July 1990 AFRC90 Paper #1 Measurement of Correlated Droplet Size and Velocity Statistics, Size Distribution, and Volume Flux in a Steady Spray Flamet C.F. Edwards Combustion Research Facility Sandia Nalionai Laboratories Livermore, CA 94551 -0969 USA R.C. Rudoff, W.O. Bachalo Aerometrics, Inc. 894 Ross Dr., Unit 105 Sunnyvale, CA 94089 USA ABSTRACT This paper reports data for the size-classified spray velocities, gasphase velocities, spray size distributions, and liquid volume flux distributions of a swirl-stabilized kerosene spray flame in a research furnace. These data were generated as part of a database to be used in validation of numerical models for steady spray flames. INTRODUCTION Development of a predictive modeling capability for two-phase furnace flames requires suitable data for comparison and validation. While many studies of spray combustion have been reported in the literature, only recently have suitable diagnostics been available for the systematic characterization of realistic spray flames. Foremost among these new diagnostics are laser Doppler velocimetry for measurement of the gas-phase flowfield and phase Doppler anemometry for simultaneous characterization of droplet size and velocity. These diagnostics make it possible to not only characterize the individual phases of complex flows, but to begin to analyze the intricate coupling that takes place between them. This paper reports the results of a study to characterize both the gas and condensed phases of a swirl-stabilized spray flame. The Phase Doppler Particle Analyzer (PDPA) is used so that the aerodynamics of the flow may be separated by droplet size class. The data reported herein form part of a database which is being used for validation of numerical approaches (e.g., Marx and Edwards, 1990). Previous characterizations of spray flames have been performed lIsing the phase Doppler method. In 1986, Mao et al., reported results from both a non-reacting and burning spray. These measurements were made in an open flame, and without detailed knowledge of the gas-phase flowfield, making them difficult to use for model validation. but nonetheless yielding great insight into the process of spray combustion. Samuelson and co-workers (McDonell et al., 1986, McDonell and Samuelson, 1988, 1989, Cameron et aI., 1988) have performed a series of systematic investigations of both spray behavior and the phase Doppler technique itself. Their studies include both open spray flames (or ducted flames with unregulated stoichiometry) and highly confmed spray flames (characteristic of gas turbine conditions). In these studies, both gas and spray behavior is reported. making them well suited for modeling purposes. Rudoff et al. (1989) have studied both the non-reacting and reacting flows from an unconfmed tThis work was supported by the U.S. Department of Energy, Energy Conversion and Utilization Technologies Program. Aerometrics would like to acknowledge the support of NASA lewis Research Center, Contract NAS3-25204. Ms. Valerie Lyons contract monitor. commercial oil burner. Their data demonstrate the power of the phase Doppler technique by virtue of its ability to provide both timeand size-resolved condensed-phase SLatistics. Hardalupas et al. (1989) have also used the phase Doppler method to study reacting sprays. They too performed their studies using an unconfined swirl burner. The present study differs from each of those reported above in at least one of three ways. First, two-component phase Doppler is used, and all three components of size-classified velocity are reported. Second, the flame is confined in a nearly axisymmetric research furnace, providing both well-defined boundary conditions and control of the overall air-fuel ratio. The degree of confinement is not so severe that it dictates the structure of the flow field (as with the gas turbine type of combustor) but is more akin to that found in industrial furnaces. Last, all three components of the gas-phase velocity statistics are reported along with the condensed-phase information. What this study lacks in comparison with those in the litcrature, is precise determination of the spray number dcnsity. Although volume flux could be measured with reasonable confidence and repeatability, this is not the case for number dcnsity. Unlike volume flux, which requires only an accurate measurement of the larger droplets in the flow, the number density requires that, in addition, the small droplets be accounted for with high accuracy. In the present case where a flame structure representative of that used in industry was sought using a realistic fuel (kerosene), the resulting luminous flame (and consequently high soot volume fraction) caused a fluctuating signal-lo-noise ratio that prohibitcd an accurate accounting of the smallest droplets. APPARATUS The optical access research furnace used in this study is the same as described in previous work (Edwards, 1988). Table I summarizes the operating conditions of this study, while Fig. 1 shows the significant features of the furnace . The furnace is octagonal in cross section, providing a compromise between the experimental requirement of planar window surfaces for optical diagnostics and the need to provide a near-axisymmetric cross section for modeling. The atomizer employcd in this study is a modified form of the Parker Hannifin Research Simplex-Air (RSA) atomizer which produces a hollow cone .spray with a nominal included angle of 60·. This nozzle uses swirl in both the atomizing air and fuel streams. Atomizing air swirl is impartcd by a 45" vanctype swirler just before the exit of the noz7.le. Fuel swirl is imparted by tangential entry upstream of the fuel filming surface. |