基于改进奇异值分解滤波和谱峭度的滚动轴承故障诊断

针对含噪信号的有效奇异值个数难以确定的问题，提出了一种改进的奇异值分解降噪方法--奇异值累积法。该方法通过计算奇异值的实际下降值与奇异值平均下降速度累积量的差值，并取该差值最大值点的位置作为有效奇异值的分界点来确定有效奇异值的个数。在此基础上，提出了一种基于奇异值累积法与快速谱峭度的滚动轴承故障诊断方法。采用奇异值累积法对原信号进行降噪处理，然后利用快速谱峭度确定滤波器中心频率及带宽，通过分析频段包络谱中明显的频率成分来诊断故障。该方法可以有效去除信号中的噪声，使得到的峭度值所反映的故障冲击更接近实际情况。对含内圈、外圈故障的滚动轴承实验数据进行分析，实验结果表明，相比快速谱峭度的故障诊断方法，该方法具有更好的故障识别效果。     

Eight-node conforming straight-side quadrilateral element with high-order completeness (QH8-C1)

A new eight-node conforming quadrilateral element with high-order completeness, denoted as QH8-C1, is proposed in this article. First, expressions for the interpolation displacement function satisfying the requirements for high-order completeness in the global coordinate system are constructed. Second, the displacement function expression in global coordinates is transformed into isoparametric coordinates, and the relationships between the two series of coefficients for the two kinds of displacement function expressions are found. Third, the displacement function expression is modified to satisfy the requirements of nodal freedom and interelement boundary continuity. The key to the new element construction is the derivation of the linear relationship expressions among 12 coefficients of element displacement interpolation polynomials in the global and isoparametric coordinate systems. As a result, the relationship between quadratic completeness and interelement continuity is explicitly given, and a proof of the completeness and the continuity was conducted to theoretically guarantee the validity of the derivation results. Furthermore, in order to verify the correctness of the theoretical work, nine numerical examples were performed. The computation results from these examples demonstrate that QH8-C1 exhibited excellent performance, including high simulation accuracy, fast convergence, insensitivity to mesh distortion, and monotonic convergence.     

综合改进奇异谱分解和奇异值分解的齿轮故障特征提取方法

针对齿轮故障特征微弱，在强背景噪声下难以有效提取的问题，提出了一种改进奇异谱分解（ISSD）结合奇异值分解（SVD）的齿轮故障特征提取方法。针对奇异谱分解（SSD）算法中模态参数需凭经验选取的缺陷，基于散布熵优化算法对SSD算法进行了改进，在得到既定的一组奇异谱分量的基础上，根据峭度值最大准则筛选出了最佳奇异谱分量并进行了SVD处理，采用奇异值能量标准谱自适应地确定了信号重构阶数以还原信号和提高降噪效果。最后对信号进行包络解调以提取齿轮故障特征，将所提方法运用到仿真信号和齿轮实测信号中，并同传统包络谱、SSD包络谱以及经验模态分解结合SVD(EMD-SVD)方法进行了对比分析，结果表明，所提方法的降噪和特征提取效果更佳，能够更加有效地实现齿轮故障的判别。     

桩–土–斜拉桥动力相互作用体系振动反应特性试验研究

大跨斜拉桥结构自振频率和阻尼较低,其地震响应可能受桩基础和场地土特性的影响较大,然而目前为止,由于试验条件和技术所限,尚缺乏相关的包括桩基础、场地土和上部结构在内的全模型振动台试验研究。以一座试设计的主跨1400 m超大跨斜拉桥为原型,设计并完成了一座几何相似比为1/70,且包括群桩、人工土和上部结构在内的试验模型,采用多点振动台试验技术,研究了不同加速度峰值和不同频率成分地震作用下桩–土–斜拉桥动力相互作用体系的振动反应特性。试验结果表明:桩–土–结构相互作用对斜拉桥地震响应产生影响,其影响程度与地震输入频谱特性密切相关;在纵向一致激励下,桩–土–结构相互作用受地震动加速度峰值的影响不明显,在横向一致激励下,桩–土–结构相互作用随地震动加速度峰值的增大而减小;主塔高阶振型对其地震响应的贡献明显;地震输入频谱特性影响桩–土–斜拉桥动力相互作用体系的地震响应,特别是在具有丰富长周期成分的Mexico City波作用下主梁竖向地震响应显著增大。     

基于分数阶Fourier变换累积量的目标检测算法

传统的信号检测和噪声分离方法只限于在时域或是频域进行,如果信号和背景噪声有很强的时频耦合,就难以在时频域得到好的检测结果。对此提出了分数阶Fourier变换累积量算法,该算法具有解除时频耦合和抑制噪声的特性,当选择合适的旋转角度使之与处理信号相匹配,就可以在分数阶Fourier域获得较好的检测结果。仿真结果表明,分数阶Fourier累积量算法的检测性能优于分数阶Fourier变换和4阶累积量算法,展现了分数阶Fourier变换累积量算法有优越的信号处理性能。     

基于高阶累积量短波通信电台细微特征提取与分析

首先介绍了高阶统计量基本概念,分析了通信电台的高阶统计量特征,并对几种常见短波通信信号的高阶矩和高阶累积量进行了计算分析;仿真结果表明：高阶统计量可以作为一类有效的特征参数,适用于短波通信信号特征识别研究.     

协同过滤算法在电影推荐中的应用

推荐系统是很多网站最关心的机器学习应用,因为其准确率的提高对网站收入有直接贡献。构建了一个电影推荐系统,使用基于相似度的KNN算法、Baseline预测、随机梯度下降以及SVD共4种方法进行预测评分。使用RMSE评价标准,对比了不同算法预测精度的差异和不同参数设定下预测精度的变化。     

Thermal buckling and postbuckling analysis of piezoelectric FGM beams based on high-order shear deformation theory

This article deals with the thermal buckling and postbuckling of functionally graded material (FGM) beams with surface-bonded piezoelectric actuators based on physical neutral surface concept and high-order shear deformation theory including von Kármán strain–displacement relationships. The beams are exposed to a uniform temperature field and electric field, the material properties of FGM layers are temperature-dependent and vary in the thickness direction. The approximate solutions of piezoelectric FGM beams for thermal buckling and postbuckling are obtained by a two-step perturbation method, meanwhile, the analytical solutions of Timoshenko beam model and Euler beam model are also presented. The validity of the present work can be confirmed by comparisons with previous results. The effects of the applied actuator voltage, beam geometry as well as volume fraction index of FGM beam on the critical buckling temperature, and postbuckling load–deflection relationships are investigated.     

Models for circular data from time series spectra

Circular data are those for which the natural support is the unit circle and its toroidal extensions. Numerous constructions have been proposed which can be used to generate models for such data. We propose a new, very general, one based on the normalization of the spectra of complex-valued stationary processes. As illustrations of the new construction's application, we study models for univariate circular data obtained from the spectra of autoregressive moving average models and relate them to existing models in the literature. We also propose and investigate multivariate circular models obtained from the high-order spectra of stationary stochastic processes generated using linear filtering with an autoregressive moving average response function. A new family of distributions for a Markov process on the circle is also introduced. Results for asymptotically optimal inference for dependent observations on the circle are presented which provide a new paradigm for inference with circular models. The application of one of the new families of spectra-generated models is illustrated in an analysis of wind direction data.