利用子波分析对平壁湍流猝发现象的研究

利用槽道湍流直接数值模拟的数据库,采用子波分析的方法。对平壁湍流猝发现象的多尺度特性进行了研究,在不同惊讶上对猝发平均周期进行了统计,并利用局部标度指数研究了猝发过程的奇异性。

STUDY OF BURST PHENOMENA IN WALL TURBULENCE BY WAVELET ANALYSIS

Scale multiplicity is one of the most important characters of turbulence coherent structures. In wall turbulence, burst process is a typical coherent motion in near-wall region, which plays an important role in turbulence production and transportation. Wavelet analysis is a local correlation analysis method and the localizability at both scale and position makes it very competitive in the study of turbulence coherent structures. In the present study, the time series of velocity fluctuations at 3 fixed points in the direct-numerical-simulated channel turbulence is analyzed. By wavelet analysis the signals of velocity fluctuations and instantaneous Reynolds stress are multi-scale decomposed, and the temporal scale multiplicity of these signals is studied. The socalled "spike"-events which are characterized by strong normal velocity fluctuation in viscous sublayer are analyzed at different scales. It is shown that the characteristic temporal scale of the "spike" events is smaller than T+ = 55. By studying the probability density functions, the skewnesses and flatnesses factors of transverse velocity fluctuations at different scales, it is found that the signals at smaller scales are much more intermittent than the signals at lager scales. The quadrant definition of ejection and sweep is applied to the multi-scale-decomposed signals of velocity fluctuations and the maximum life span and the mean period of burst events at different scales are obtained by conditional averaging. It is found that they both increase with scales, but they are independent of the distance to the wall. The mean burst period obtained by wavelet analysis is consistent with the results obtained by conditional sampling methods if the scale of maximum turbulent kinetic energy is regarded as the characteristic scale of burst. The local scaling exponents of the multi-scale signals of the velocity fluctuations are obtained by wavelet coefficients, and the mean period of negative values of the local scaling exponents at different scales are calculated, which are consistent with the mean burst period. It indicates that the burst procedures are closely related to the singularities of the velocity fluctuations.