壁湍流相干结构的辨识

本文评价了辨识湍流边界层近壁区域相干结构的条件采样法、自相关函数法、数字滤波法和子波分析法。结果表明：条件采样方法对门限值和平均窗宽等检测参数具有较强的依赖性，可以通过调整这些参数来选择最佳的参数值使得相干结构的误判概率达到最小。自相关函数法不需要任何参数就能够得出确定的结论，因此是一种比较客观的方法。但是，这种方法只能从统计上确定平均猝发周期，并不能实时地辨识相干结构。数字滤波法能够将湍流脉动信号分解为大尺度涡和近似各向同性的小尺度涡两部分，但是如何去掉大尺度信号中的非相干部分是仍然需要解决的问题。子波分析是分解湍流信号辨识湍流相干结构的有效手段，它能够将湍流信号在时域空间和频域空间进行双局部化分解，用能量最大准则确定相干结构对应的时间尺度参数，用子波逆变换提取相干结构对应的速度信号波形

The Identification of Coherent Structure in Wall Turbulence

An evaluation of the effectiveness of conditional sampling,auto correlation function,digital low pass filter and wavelet analysis coherent structure identification algorithms has been done while they are used to detect the coherent structures that arise in the near wall region of turbulent boundary layer.The conditional sampling methods were found to be highly dependent on the operational parameters,i.e.threshold levels and averaging or window times.These parameters needs to be adjusted so that the optimum parameter is set and the probability of a false detection is minimized.The auto correlation function method never needs any parameter so it is objective.But it can only determine the average burst period statistically and can not detect the coherent structure instantaneously.The digital low pass filter method make it possible to decompose the turbulent fluctuating signal into two parts:one is the large scale eddies in which the coherent structures are expected to exist and the other is the small scale eddies that are nearly isotropy.But how to get rid of the noncoherent signal from the large scale part is still a problem to be solved.Wavelet analysis is a very promising tool for decomposing turbulent signals and identifying coherent structures in turbulence.It can perform the double local decomposition of the turbulent signals both in time space and in frequency space.The energy maxima critera based on the wavelet coefficients can determine the time scale parameter that is related to the coherent structures.The signal that corresponds to coherent structures is reconstructed by the formula of the inverse wavelet transformation.