声学多普勒流速仪测量湍流流速的峰值降噪方法

声学多普勒流速仪常用于测量水流三维瞬时流速,受外界以及仪器自身等多种因素影响,采集的流速时间序列中不免存在噪声干扰。为了获取真实反映湍流瞬时紊动过程中的高频流速信号,需对流速仪测量流速信号中含有的噪声值进行过滤,为此提出了一套流速信号预处理方法。先对大于3.5倍样本标准差的流速数据进行修正;其次分析样本中局部区域数据的变化情况,对大于2倍局部样本标准差的数据进行插值修正;根据傅里叶线性变换性质,提出从流速信号功率谱中过滤掉仪器固有噪声功率谱。水槽实验数据分析结果表明:调整数据量不超过总数的12%,符合柯尔莫戈洛夫的-5/3次幂频率增加规律,对高频噪声降噪作用明显,为湍流瞬时紊动过程分析提供了更为准确的方法。

Study on despiking method of turbulent flow velocity measured by ADV

ADV was usually used to measure three-dimensional momentary velocities of turbulent flow. However, this method is sensitive to external environment and instrument itself, which can lead to spike noise in collecting velocity time series. In order to obtain the true and high-frequency velocity signals of instantaneous turbulent flow, a new velocity signal pretreatment method should be put forward to filtrate the noise signal. Firstly the velocity data that is 3. 5 times greater than sample standard deviation was revised;then the data signal variation of local area in sample is analyzed and furthermore the velocity data that is 2 times greater than sample standard deviation was reversed by interpolation method. So the power spectrum filtering method for filtrating instrument noise was put forward based on Fourier linear transformation. The analysis results of experimental data shows that the number of mended data is less than 12% of the total;the velocity power spectrum of processed data is more aligned with Kolmog-orov power spectrum. The proposed method has an obvious effect in despiking of high frequency noise, which provides more accu-rate analysis basis for turbulence characteristics analysis.