Statistical Interpretation of Power Spectral Densities Measured by Picosecond Time-Resolved Laser-Induced Fluorescence in Turbulent Nonpremixed Flames

摘要

Picosecond time-resolved laser-induced fluorescence (PITLIF) has been applied to time-series measurements of OH, CH, and number density in turbulent nonpremixed flames with the intention of better understanding scalar fluctuation rates. Power spectral densities and autocorrelation functions were computed from the time series and then used to calculate integral time scales for many axial and radial locations in seven jet diffusion flames. The OH autocorrelation functions were found to collapse when normalized by their integral time scales. A stochastic model based on the laminar flamelet concept was developed to predict these integral time scales. Improvements over a prior version of this model and a more systematic application to H2/CH4/N2 flames permitted a reasonable prediction of the experimental trends. CH autocorrelation functions were also found to be self-similar except at locations close to the jet centerline or for Reynolds numbers approaching blowoff. Assumptions required for the time-series simulations were assessed via time-series measurements of total number density in the same flames, and the required input data were extracted by using time-averaged measurements from other research groups. Additional OH measurements in an H2/N2 jet flame provided the first direct comparison to integral time scales predicted from large-eddy simulations (LES). The predicted temporal scales were found to be a factor of two lower than the scalar measurements, but the LES predictions reproduced the correct shape for the OH power spectral density.