基于主元分析与动态时间弯曲的故障诊断方法及应用研究

轴承是工程实际中常用而又极易损坏的部件,特别是对其早期微弱响应的辨识,具有重要的社会价值和意义。为提高运转轴承的安全可靠性和可维护性,提出了基于主元分析与动态时间弯曲距离的故障诊断方法,它可以准确对早期微弱动态响应辨识、诊断。该方法首先将典型故障样本信号与待测信号小波去噪并EMD分解,并对若干固有模态分量主元分析求取主元,然后对主元分量进行分析,获得相关特征值组成特征向量,计算待测信号与已知故障样本信号特征向量的弯曲距离,弯曲距离越小表明两信号越相似,从而辨识故障。此外,还可将其应用于转子、碰磨、齿轮故障诊断中,工程应用实例表明该方法可以准确故障分类,高效故障诊断。     

基于经验模态分解和小波阈值的冲击信号去噪

冲击信号是非线性的并且容易受到噪声污染。为研究冲击信号去噪的问题,本文针对经验模态分解(Empirical Mode Decomposition,EMD)去噪和小波阈值去噪方法存在的不足,提出了基于EMD的小波阈值去噪方法。单纯的EMD去噪方法会在去除高频噪声的同时压制高频的有效信息。本文将EMD与小波阈值去噪相结合,利用连续均方误差准则确定含噪较多的高频固有模态函数(Intrinsic Mode Function, IMF),对高频IMF分量进行小波阈值去噪,以分离并保留这些分量中的有效信息,同时保持低频IMF分量不变。对模拟数据和实际冲击信号进行去噪处理,结果表明,基于EMD的小波阈值去噪方法的去噪效果优于单纯的EMD去噪方法和小波阈值去噪方法。     

基于EMD的拉曼光谱去噪方法研究

经验模态分解(EMD)方法是一个以信号极值特征尺度为度量的时空滤波过程,它充分保留了信号本身的非线性和非平稳特征,在信号的滤波和去噪中具有较大的优势。文章在介绍EMD分解方法的基础上,结合EMD的多尺度滤波特性,提出了一种新的拉曼光谱去噪方法——EMD阈值去噪法。该方法首先对含噪的拉曼光谱信号做EMD分解,得到各阶本征模态函数(IMF),然后对高频的IMF分量用阈值法进行处理,把经过阈值处理后的高频IMF分量与低频IMF分量叠加得到重构的信号,即去噪信号。通过处理对二甲苯的拉曼光谱信号,分析了在不同噪声水平上不同去噪方法的处理效果。实验结果表明EMD阈值去噪法有效地去除了噪声,较好地保留了光谱的细节信息,与小波阈值去噪方法相比较具有自适应的优势,在拉曼光谱去噪中有很好的应用前景。     

基于峰值信噪比和小波方向特性的图像奇异值去噪技术

提出一种利用小波变换子图像不同的方向特性和峰值信噪比进行奇异值分解的图像去噪算法。由于图像经过小波变换后,低频子图像集中了原图像的大部分能量噪声,故仅作简单维纳滤波;而噪声则主要集中在小波域中的三个不同方向的高频子图中,且系数较小,因此可以利用奇异值分解进行去噪处理,即用较大的奇异值和对应的特征向量重构出去噪图像,然而由于奇异值分解固有的行列方向性,对于高频对角线子图重构出的图像去噪效果不理想,故采取旋转至行列方向后再进行常用的奇异值滤波;最后将去噪后的低频和高频子图进行小波反变换重构出最终的去噪图像,其中重构所需的奇异值个数由图像的峰值信噪比确定。 实验结果表明,该方法在有效去噪的同时较好的保留了原有的高频细节信息。     

变分模态分解-排列熵方法用于分布式光纤振动传感系统去噪

提出一种变分模态分解-排列熵的去噪方法,分析并设定排列熵中关键参数和阈值,进而通过排列熵来确定变分模态分解的分解层数值,将分解的各模态进行重构以实现对振动信号的去噪。通过仿真测试来验证该方法在正交性、完备性、信噪比和效率方面的优越性,最后对系统采集的实际振动信号进行去噪处理。实验结果表明,与现有的经验模态分解-相关系数和完全经验模态分解-相关系数方法相比,所提方法对触网、车轮碾压和雨淋三种振动信号具有最优的去噪信噪比(含噪信号与降噪值之比),分别为32.5358 dB、30.5546 dB和29.3435 dB,耗时也较少,分别为1.4432,1.6320,1.2349 s,信号模式识别准确率最高,均在99%以上。     

超声辅助加工系统的刀具状态自感知算法

超声波振动台内含压电材料,可以拾取切削过程产生的振动信号,实现不借助外部传感器刀具工作状态的自感知。为了从刀具振动信号中获取有效信息,该文提出一种基于经验模态分解的时频域重构算法。首先,采用经验模态分解算法将原始信号分解,得到多个固有模态函数分量和残差分量;其次,计算原始信号与各分量之间的时频域互相关系数;再次,归一化时频域互相关系数作为权重值,将固有模态函数分量和残差进行重构;最后,通过数值仿真和超声辅助加工实验,验证了基于经验模态分解的时频域重构算法的去噪性能,提取了信噪比为5.03 dB的目标信号,从而实现了超声辅助加工系统的自感知功能。     

基于LCEEMD的低信噪比拉曼光谱自适应去噪方法研究

在生物体拉曼光谱快速采集或低功率采集过程中,往往会获得低信噪比拉曼光谱。针对低信噪比光谱数据,提出应用补充总体经验模态方法(CEEMD)分解拉曼光谱,并且依据特征模态分量的归一化排列熵值(NPE)按比例扣除噪声成分的方法,称为局部补充总体均值经验模分解方法(LCEEMD)。LCEEMD方法不仅解决了经验模态(EMD)分解中高频信号与噪声的模态混叠问题,还有效降低了总体经验模态分解法(EEMD)中的残留噪声。仿真数据实验显示,LCEEMD方法在处理10db信噪比模拟光谱时获得了39.615 0 db信噪比,0.001 17标准差和0.999 9相关系数。在人体皮肤拉曼光谱试验中,LCEEMD方法滤波后数据准确呈现出角质层脂质酰胺I带激发拉曼强谱峰以及甘油三酸酯中(CO)酯微弱谱峰。在水稻叶片可溶性糖定量预测模型中,LCEEMD方法取得了0.871 7预测相关系数和0.912 0预测标准误差,优于EMD和EEMD软阈值去噪(0.511 4,1.647 8和0.638 2,1.508 8)。LCEEMD方法实施过程中,根据去噪性能指标反馈调整归一化排列熵阈值,直至获得最佳去噪效果,滤波过程无需参数设置,可以自适应实现。     

基于滑动峭度相关性准则的局部特征尺度分解分量筛选方法

轴承故障振动信号具有非平稳、非线性特征,且可视为多个调幅-调频分量的叠加,单分量的包络蕴含了轴承的故障特征。局部特征尺度分解可将振动信号准确分解为多个内禀尺度分量之和,某些分量能清晰反映轴承的运行状态,根据包络谱可进行故障诊断。为了准确筛选有用分量,提出了基于滑动峭度相关性准则的分量筛选方法。首先,对信号进行局部特征尺度分解,得到若干个内禀尺度分量;然后,对分量和原始信号分别计算滑动峭度,生成时间序列;最后,依据分量滑动峭度序列与原始信号滑动峭度序列的互相关系数筛选有用分量。通过轴承内圈故障数据分析发现：有用分量与非有用分量之间的滑动峭度互相关系数比互相关系数差异明显,区分度更大,有益于分量的分类、筛选。     

基于张量的平稳小波变换红外图像去噪

提出了一种基于张量的平稳小波变换红外图像去噪方法.采用平稳小波对噪声红外图像进行分解,保持低频近似图像不变,将所有尺度上的水平、垂直和对角方向的高频细节图像组合为一个立方体,形成三阶张量,通过多线性代数方法估计信号小波系数,这种处理方式没有破坏小波系数之间的固有空间关系,同时考虑到了尺度问和尺度内小波系数的相关性,优于传统的基于线性最小均方误差的信号小波系数估计算法,最后由低频近似图像与估计的高频细节图像通过平稳小波逆变换得到去噪图像.实验结果表明,该方法在性能指标和视觉质量上优于传统的平稳小波域最小均方误差去噪算法,为小波系数的较准确估计提供了一种全新思路.     

激光雷达信号的可变间隔阈值经验模式分解去噪法

针对采用经验模式分解直接阈值(EMD-DT)和经验模式分解间隔阈值(EMD-IT)在激光雷达回波信号的去噪应用中会产生的模态混叠现象,采用一种可变间隔阈值的经验模式分解(EMD-SIT)的去噪方法。首先,对信号进行经验模式分解。然后,采用过零率方法将分解出的含有噪声的固有模态函数分离。最后,应用过零点阈值,设立一个新的可变阈值,将EMD-IT和EMD-DT有效融合对信号进行去噪。通过与多种阈值的仿真对比以及激光雷达的回波信号去噪实验,结果表明该方法可以有效地去除噪声,抑制模态混叠,较EMD-IT和EMD-DT更具有优越性,因此有着很好的应用前景。     

白光和红绿蓝光LED可调光源研制

为了提高LED光源色温和亮度的调节精度和准确度,结合色温由低向高变化时光色所呈现的渐变特点,提出了一种低色温白光LED灯珠、高色温白光LED灯珠加红绿蓝光LED灯珠补偿式调光的方法.将色温分成三个部分进行调节,每个部分选用不同的LED灯珠组合来进行调光.实验结果表明:不同组合情况下的LED光源的初始输出色温相对于目标值的偏差范围在1％以内;亮度可以在保证色温不变的情况下独立进行调节,初始输出值与目标值的偏差范围在1％以内;经过微调之后可以达到目标值;达到了色温和亮度独立调节的要求;光源发光稳定,不会因为长时间工作而影响调节精度.     

Muonic and pionic L- and M-series in carbon and oxygen and the pionic 2p level shift and width of oxygen

Muonic and pionic X-rays of the L- and M-series in C and O have been measured with a Si(Li) detector in the energy range between 7 keV and 60 keV. The target consisted of mylar (C₅H₄O₂). Energies and intensities of 21 transitions have been determined. The strong interaction shift of the pionic 2p level in O was measured and found to be +4.1 ±2.3 eV. The measured width of this level is 11±6 eV. The measured yields have been compared with cascade calculations.     

Visualization and interferometry for cylindrical and spherical sample tomography and profilometry

A technique combining image processing and laser interferometry for visualizing and detecting the deformation of transparent cylindrical and spherical sample is proposed. This deformation includes geometric deformation such as volume transition in profilometry and physical deformation such as refractive index change in tomography. Phase contour lines are used for quantitative analysis and graphical representation of the deformation. This method allows us to visually detect the spatial variation of the deformation field and to evaluate the test quality such as misalignment of optical system. A theoretical analysis using phase contour map to characterize the deformation field is described in detail. A method using phase contour map to qualify the interferometric test is proposed. Analysis of test examples is carried out. Suggestions on using phase contour line method to ameliorate test system design are finally discussed.     

Nonlocal potentials and their exact and approximate local and velocity-dependent equivalents

We show that good approximations to the exact equivalent local potential (ELP) and damping factor of a nonlocal Perey-Buck potential can be calculated in the partial wave WKB approximation of Horiuchi. The exact ELP and damping factor are obtained by means of a method previously given by one of us. We also confirm that an approximate ELP proposed by Bauhoff et al. is of comparable accuracy as the Horiuchi approximation. Thesel-dependent ELP's exhibit reduced attraction in the interior and provide a test for higher order WKB approximations. We subsequently obtain an equivalent velocity dependent potential (EVDP) which is even exactly wave function equivalent to the original nonlocal potential. This almost local potential, unlike the trivial equivalent local potential, is smooth and well-behaved and is therefore particularly useful in nuclear reactions where the off-shell behaviour of the potential is important.     

Spin and Statistics and First Principles

It was shown in the early seventies that, in Local Quantum Theory (that is the most general formulation of Quantum Field Theory, if we leave out only the unknown scenario of Quantum Gravity) the notion of Statistics can be grounded solely on the local observable quantities (without assuming neither the commutation relations nor even the existence of unobservable charged field operators); one finds that only the well known (para)statistics of Bose/Fermi type are allowed by the key principle of local commutativity of observables. In this frame it was possible to formulate and prove the Spin and Statistics Theorem purely on the basis of First Principles.     

Quantum and Classical Disparity and Accord

Discrepancies and accords between quantum (QM) and classical mechanics (CM) related to expectation values and periods are generally found for both the harmonic oscillator (SHO) and a free particle in a box (FPB), which may apply generally. These indicate non-locality is expected throughout QM. The FPB energy states violate the Correspondence Principle. Previously unexpected accords are found and proven that 〈x ²〉 _CM =〈x ²〉 _QM and τ _CM =τ _QMb (beat period i.e. beats between the phases for adjoining energy states) for the SHO for all quantum numbers, n. However, for the FPB the beat periods differ at small n. It is shown that a particle’s velocity in an infinite square well varies, no matter how wide the box, nor how far the particle is from the walls. The quantum free particle variances share an indirect commonality with the Aharonov-Bohm and Aharonov-Casher effects in that there is a quantum action in the absence of a force. The concept of an “Expectation Value over a Partial Well Width” is introduced. This paper raises the question as to whether these inconsistencies are undetectable, or can be empirically ascertained. These inherent variances may need to be fixed, or nature is manifestly more non-classical than expected.