Savitzky Golay-WPT在滚刀主轴振动信号降噪中的应用

为消除滚齿机在加工过程中,滚刀主轴振动信号因环境影响而产生的噪声信号,提出一种基于Savitzky Golay -WPT的信号降噪方法。对原信号进行计算,得其最佳小波包分解树;根据最佳分解树,进行小波包变换（WPT）,得小波包系数;利用阈值函数对小波包系数进行筛选;结合最小二乘拟合方法对小波包筛选后系数进行重构。结果表明：与传统小波包和CEEMDAN相比,所提方法降噪性能分别提高3135%和2271%;在实际加工数据中,与传统小波包方法对比,该方法可减少中心频率周边干扰,使中心频率特征更突出,降噪效果更明显。     

小波变换在薄膜表面图像信号降噪中的应用

目的：当前,以原子力显微镜为代表的扫描探针显微镜设备可以获取纳米尺度薄膜样品的表面图像,但这些图像存在不同程度的噪声,影响图像质量和信息判断。为了更准确获取这些薄膜表面状态,需要对薄膜样品表面图像数据和信息进行降噪处理。方法结合AFM等设备成像特点以及小波变换的时频局域性特点,在介绍小波变换基本理论和噪声来源分析基础上,提出了一种多层小波分解去噪算法。传统的信号理论是建立在傅里叶变换基础上的,而傅里叶变换作为一种全局性的变化,其有一定的局限性,无法同时表述信号在时域和频域的局部性质,而这些局部特征恰恰是非平稳信号性质最关键的部分。小波变换保留了窗口傅里叶变换局部化的优点,改变了窗口傅里叶变换窗口函数大小固定的缺点。结果原始图像信号的频率在0 Hz到4000 Hz都有分布。通过小波变换后,信号波形更光滑,频谱在500 Hz到2000 Hz之间分布。结论将小波变换应用于薄膜表面图像信号降噪中,通过实验证明通过小波变换可以有效去除信号中的噪声部分。     

A crevice corrosion study by Electrochemical Noise Analysis

The use of the Electrochemical Noise Analysis (ENA) for the evaluation of crevice corrosion is illustrated in the case of AISI 430 stainless steel in 3% sodium chloride. A crevice former was used in order to induce a crevice corrosion attack. Current and potential noise signals were simultaneously recorded allowing the determination of the noise resistance (R_n). ENA was able to detect the four stages mechanism of crevice corrosion. The comparison of R_n with the Polarization Resistance (R_p) determined by Electrochemical Impedance Spectroscopy (EIS) gave good agreement in particular during the initiation and propagation of the attack. The evaluation of the noise data in the frequency domain gave interesting results in particular in the evaluation of the roll‐off slopes in the Power Spectrum Density (PSD) plot that are correlated to the corrosive status. Finally a new analysis for noise data is shown. The application of spectral ratio discriminant function to noise data in the frequency domain, permits to deduce the best sampling frequency and sampling duration for ENA acquisition, able to discriminate between two different situations.     

Ultrasonic NDE of composite material structures using wavelet coefficients

A wavelet-based method is proposed to perform the analysis of NDE ultrasonic signals received during the inspection of reinforced composite materials. The non-homogenous nature of such materials induces a very high level of structural noise which greatly complicates the interpretation of the NDE signals. By combining the time domain and the classical Fourier analysis, the wavelet transform provides simultaneously spectral representation and temporal order of the signal decomposition components. To construct a C-scan image from the wavelet transform of the A-scan signals, we propose a selection process of the wavelet coefficients, followed by an interpretation procedure based on a windowing process in the time–frequency domain. The proposed NDE method is tested on cryogenic glass/epoxy hydrogen reservoir samples.     

基于小波阈值法和改进移动平均法的电梯钢丝绳断丝损伤信号去噪方法研究北大核心

针对使用传统方法进行电梯钢丝绳断丝损伤信号降噪过程中存在Gibbs效应,去噪后信噪比较低的问题,提出一种基于小波阈值法和改进移动平均法的降噪方法。对实测信号进行小波变换多尺度分解,得到不同的尺度系数;将各尺度系数与确定的固定阈值比较,去除小于该阈值的尺度系数,保留大于该阈值的尺度系数;对保留下的各尺度系数重构,得到初步降噪后的信号;运用改进移动平均法对其作平滑处理,得到最终信号。结果表明:该方法能够有效抑制Gibbs现象,具有很好的降噪效果,而且信噪比很高,有助于钢丝绳断丝故障的定量识别和特征提取。     

基于小波分解和BP神经网络的磁巴克豪森噪声信号分层分析研究

因小波变换具有多尺度分析的特点,在时频两域都有表征信号局部特征的能力,因此采用小波分解方法研究了不同时频成分的磁巴克豪森（MBN）信号随温度和应力变化的灵敏度问题。采用db5小波对MBN信号进行6层小波分解,提取各层分解系数的均值和均方根,并讨论分析了各特征值随所加应力以及温度变化的相对变化关系。研究表明,在试样的弹性范围内,低频系数和各层高频系数的均值和均方根都随压应力的增加而减小;各层高频系数的均值和均方根随温度的升高而降低,低频系数的均值和均方根随温度的升高而升高。最后将温度、原始MBN信号以及各分解系数的均值和均方根作为神经网络的输入,压应力作为其输出建立神经网络模型,结果表明该神经网络模型与之前没有用小波分解时的神经网络模型相比,检测应力的准确性更高。     

Correlation between electrochemical impedance and noise measurements of waterborne coatings

A modified equivalent electric circuit is proposed to establish correspondence between impedance and noise measurements in waterborne coatings in saline media (NaCl). The polarization resistance (R_p) given by the impedance modulus when frequency approaches to zero can not directly be compared with noise measurements since this latter accounts only for the pore resistance of the paint of low barrier protective coatings while R_p parameter represents the sum of all the resistances at the system. The results have shown that the pore resistance data can be attributed to noise resistance when (i) the coating exhibit porosity presenting diffusion mechanism as rate-determining step, and (ii) when the sampling frequency of recorded data gives a Nyquist frequency of 1 Hz.     

基于小波包变换与自适应阈值的SMT焊点图像去噪

为提高表面组装技术焊点图像去噪效果,提出一种基于小波包变换和自适应阈值的焊点图像噪声去除新方法.利用小波包变换对图像进行多层分解,通过对图像小波包树系数的分析,对小波包树系数高频部分和低频部分进行Wiener滤波;为提高去噪性能,通过计算小波包系数对应的中值绝对方差估计,提出一种改进的自适应小波包阀值算法,并对图像进行...     

Application of the wavelet transform in mechanical spectroscopy and in Barkhausen noise analysis

In this work, the wavelet transform and the Fast Fourier Transform FFT are introduced to analyze the quality of the strain response signal in order to improve the accuracy of computation of both the logarithmic decrement and the mechanical loss angle. The wavelet transform of the strain response signal yields a three-dimensional representation (the time-scale joint representation), i.e. space (time), frequency (scale) and amplitude. To emphasize that not only time but also frequency content of the strain response signal collected from a mechanical spectrometer is identified, the author suggests that the time-scale joint representation of the strain response signal should be called ‘Identified Strain Response Signal — ISRS’. The wavelet transform was shown to be an excellent tool to test the quality of the strain and/or of the stress signal in a mechanical spectrometer working in a resonant or subresonant mode. A new approach to the non-stationary Barkhausen noise (BN) signal based on the wavelet transform is also presented. The BN level is usually expressed with the magnetoelastic parameter MP which is a relative number proportional to the root mean square level of the BN. It was found that the MP value is not a reliable parameter for the identification of stress concentrations in ferromagnetic materials. Indeed, a stress concentration resulting from internal stresses can be successfully revealed by the time-scale joint representation of the BN signal. The author suggests that the wavelet transform of the non-stationary RN signal should be called the ‘Identified Barkhausen Noise — IBN’. This approach produces a physically reliable relationship between the IBN and either the level of internal stresses or fine variations in the microstructure.     

基于小波包变换的拟合非光滑目标谱的地震波生成(英文)

地震时程分析中需要使用符合规范的目标反应谱的地震记录,而以往的拟合目标谱的方法大多是针对目标谱为光滑曲线的情况的。在许多情况下,用单一的确定的反应谱,往往需要拟合不光滑的目标谱。给出一种基于小波包变换的、能够很好地拟合非光滑目标谱的地震时程调整方法。经过观察,发现地震记录经过小波包分解后的成分的频带中心频率与小波包成分的反应谱的峰值频率存在一定的偏差。基于这个现象给出了5个选择迭代点的策略,通过两个计算实例比较了这5个策略的效果。结果表明,以小波包成分的反应谱的峰值频率作为迭代点的策略能获得很好的拟合效果,并且在迭代过程中稳定趋向收敛,能很好地适用于不光滑目标谱的拟合。     

Integrated interpretation of dual frequency induced polarization measurement based on wavelet analysis and metal factor methods

In mineral exploration, the apparent resistivity and apparent frequency (or apparent polarizability) parameters of induced polarization method are commonly utilized to describe the induced polarization anomaly. When the target geology structure is significantly complicated, these parameters would fail to reflect the nature of the anomaly source, and wrong conclusions may be obtained. A wavelet approach and a metal factor method were used to comprehensively interpret the induced polarization anomaly of complex geologic bodies in the Adi Bladia mine. Db5 wavelet basis was used to conduct two-scale decomposition and reconstruction, which effectively suppress the noise interference of greenschist facies regional metamorphism and magma intrusion, making energy concentrated and boundary problem unobservable. On the basis of that, the ore-induced anomaly was effectively extracted by the metal factor method.     

基于时频分析的奥氏体焊缝超声检测信号处理

对奥氏体焊缝超声检测信号处理方法进行了研究.提出一种基于匹配追踪和小波变换两种时频分析方法的信号处理方法,分别从能量和频率的角度区分缺陷信息和材料散射噪声,提高超声检测回波的信噪比.为了验证此方法的有效性,建立了超声检测试验系统,对53 mm厚焊缝试块进行检测试验,采集了含有大量材料噪声的实际检测信号.选用sym8小波...     

Gradual wear monitoring of turning inserts using wavelet analysis of ultrasound waves

Discrete wavelet transforms of ultrasound waves is used to measure the gradual wear of carbide inserts during turning operations. Ultrasound waves, propagating at a nominal frequency of 10 MHz, were pulsed into the cutting tools towards the cutting edge at a burst frequency of 10 KHz. The reflected waves off the mark, nose and flank surfaces were digitized at a sampling rate of 100 MHz. Daubechies Quadrature Mirror Filter pair was used to decompose ultrasound signals into frequency packets using a tree structure.Normalized signals in each level of decomposition were used to search for a neural network architecture that correlates the ultrasound measurements to the wear level on the tool. A three-layer Multi-Layer Perceptron architecture yielded the best correlation (95.9%) using the wave packets from the fourth level of decomposition with frequencies 3.75–4.375 and 5.625–6.875 MHz.     

一种改进全变差正则化的Shearlet 自适应带钢图像去噪算法

目的有效去除生产现场所采集的带钢图像上的混合噪声。方法结合Shearlet变换具有较好的稀疏表示图像特征的性质与全变分各向异性扩散的优点,提出一种带钢图像去噪新算法。对Shearlet变换分解后的图像进行硬阈值处理,再进行Shearlet变换重构形成估计图像,采用改进自适应的变差正则化的极小化迭代模型对估计图像进行迭代修正。结果去噪后的图像具有很好的视觉效果,避免了伪吉布斯效应的产生。在强噪水平下,对比新模型与小波去噪,PSNR提高了约9 d B,均方差降低了约319。结论该方法获得了较好的峰值信噪比增益,使信号幅度有较高的保真度,具有更好的平滑噪声和边缘保持功能。     

变转速变载荷工况下的齿轮智能故障诊断北大核心

针对变转速变载荷工况下的齿轮故障检测、识别和分类问题,提出一种基于最大重叠离散小波包变换和人工神经网络的智能故障诊断新方法。研究自相关谱峭度图中的最大重叠离散小波包变换,并采用它将复杂的齿轮故障振动信号分解为频带和称为节点的中心频率。推导出每个节点的平方包络的自相关,以便计算每个节点在每个分解层次上的峭度,减少了非周期性脉冲和噪声的影响。将上一步得到的特征矩阵作为径向基函数神经网络的输入,从而实现齿轮状态的自动分类。并在变转速变载荷(16种)工况下对健康状态和5种不同类型齿轮故障的齿轮箱进行了具体测试分析。结果表明:该方法可以更好地提取特征信息,为齿轮故障诊断定位合适的解调频带,提高了所有工况下齿轮故障诊断的准确率。     

基于小波变换的发动机表面缺陷图像去噪方法的研究

目的 滤除发动机表面缺陷图像上的噪声,使发动机表面缺陷信息得以更好地呈现。方法 首先利用小波变换将发动机表面缺陷含噪图像进行系数分解,获取不同的小波系数;接着利用支持向量机对小波分解系数进行分类,以达到将噪声信号与非噪声信号进行分离的效果;最后利用插值运算对硬阀值函数进行优化,以克服函数不连续性引起的振铃效应等弊端,使得去噪后图像能够保持更多的细节信息。通过实验仿真将所提方法以及中值滤波、双边滤波方法的去噪效果进行对比。结果 所提方法去噪后图像与中值滤波以及双边滤波方法去噪后图像相比,具有更高的PSNR值以及SSIM值。测试图像噪声强度为25%时,所提方法去噪后图像的PSNR值以及SSIM值较中值滤波方法去噪分别提高了20.66%以及11.89%,较双边滤波方法去噪分别提高了10.30%以及5.48%。结论 所提方法比中值滤波、双边滤波方法具有更好的去噪效果,能够对发动机表面缺陷图像的噪声进行去除,并较好地保留图像的细节信息。     

数字超声系统脉动噪声的估计与消除

采用高通滤波器和小波包变换技术对低碳钢摩擦焊接头进行超声无损检测，研究表明，高通滤波后，信噪比有所提高，但频率滤波器在滤除噪声的同时，也滤除了缺陷信息中所含的低频成分。小波包变换是一种比小波变换频率线性度更好的时频分析方法，能同时兼顾信号的陡变和缓变特征，从而清楚地区分时变信号和长时间的类周期信号。实验证明，小波包换能正地估计出脉动噪声，从而有效地去除之，提高信号的可观察性和信噪比。     

Electrochemical noise analysis of localized corrosion by wavelet transform

Electrochemical noise (EN) was measured during pitting and stress corrosion tests in NaCl solutions for type 403 stainless steel tempered at 390 °C for 2 h. The EN data were analyzed mainly by wavelet transform in terms of the energy contribution of smooth crystal and detail crystals in three frequency regions. In the early stage of the pitting tests, general corrosion occurs, and contribution of smooth crystal to the total energy was greater than that of low frequency detail crystal, but less with the development of meta-stable pitting. A similar trend was observed in the early stage of stress corrosion cracking (SCC) tests. The contribution of low frequency detail crystals to the total energy decreased as stable pits were formed, and it was less than the energy contribution of smooth crystal when crack propagation occurred in the later stage of SCC. In addition, the contribution of energy of high frequency detail crystal is shown to be indicative of the occurrence of plastic deformation. These results demonstrate that the onset of detailed processes in pitting and SCC can be sensitively detected by wavelet analysis in terms of the energy contribution of smooth crystal and low and high frequency detail crystals, but not with that based on the change in position of peak in energy contribution of detail crystals.     

Interpretation of electrochemical noise by wavelet transform and fractal analysis

Electrochemical noise(EN) was measured during pitting and stress corrosion tests in 3.5 % NaCl solutions for type 403 stainless steel tempered at 520 °C and 610 °C. The energy contributions of smooth crystal and detail crystals in three frequency divisions were analyzed after wavelet transform and compared with the change in the Hurst parameter based on the rescaled range analysis. The contribution of smooth crystal to the total energy is greater than that of low-frequency detail crystal in general corrosion and stable pit formation, but it is less in both meta-stable pitting and growth of stable pits. Crack growth rate can be estimated qualitatively by the evaluation of the energy contributions of smooth and detail crystals, and correlation was found between the change in the Hurst parameter and that in the energy contribution of smooth crystal or one of detail crystals in three frequency divisions.