基于参数优化变分模态分解和多尺度熵偏均值的行星变速箱故障特征提取

针对某型装甲车辆行星变速箱行星轮故障特征难以提取的问题,提出了结合参数优化变分模态分解(VMD)和多尺度熵偏均值的故障特征提取方法。为克服VMD算法参数选取依赖经验的弊端,采用粒子群优化算法对VMD参数进行优化。使用参数优化后的VMD算法对信号进行分解,并依据互信息选取有效分量对信号进行重构。多尺度熵能反映信号在多尺度下的复杂度,偏均值可以反映多尺度熵的均值和变化趋势。采用基于多尺度熵的综合指标多尺度偏均值,以全面反映振动信号在多尺度下的特性,用于衡量行星变速箱运行状态,从而进行故障特征提取。行星变速箱实验数据处理结果表明,新方法可以更加有效的提取行星变速箱故障。     

Imbalanced fault diagnosis of rotating machinery using autoencoder-based SuperGraph feature learning

Existing fault diagnosis methods usually assume that there are balanced training data for every machine health state. However, the collection of fault signals is very difficult and expensive, resulting in the problem of imbalanced training dataset. It will degrade the performance of fault diagnosis methods significantly. To address this problem, an imbalanced fault diagnosis of rotating machinery using autoencoder-based SuperGraph feature learning is proposed in this paper. Unsupervised autoencoder is firstly used to compress every monitoring signal into a low-dimensional vector as the node attribute in the SuperGraph. And the edge connections in the graph depend on the relationship between signals. On the basis, graph convolution is performed on the constructed SuperGraph to achieve imbalanced training dataset fault diagnosis for rotating machinery. Comprehensive experiments are conducted on a benchmarking publicized dataset and a practical experimental platform, and the results show that the proposed method can effectively achieve rotating machinery fault diagnosis towards imbalanced training dataset through graph feature learning.     

A smoothing inexact Newton method for variational inequality problems

In this paper, we reformulate the variational inequality problem as an equivalent smooth non-linear equation system by introducing the Chen–Harker–Kanzow–Smale smoothing function. A new smoothing inexact Newton algorithm is proposed to solve the smooth equations. In each iteration, the corresponding linear system is solved approximately. We prove that the proposed algorithm converges globally and superlinearly under mild conditions. Preliminary numerical results indicate that the method is effective.     

A 3D multi-grid algorithm for the Chan–Vese model of variational image segmentation

Variational segmentation models provide effective tools for image processing applications. Although existing models are continually refined to increase their capabilities, solution of such models is often a slow process, since fast methods are not immediately applicable to nonlinear problems. This paper presents an efficient multi-grid algorithm for solving the Chan–Vese model in three dimensions, generalizing our previous work on the topic in two dimensions, but this direct generalized method is low performance or unfeasible. So here, we first present two general smoothers for a nonlinear multi-grid method and then give our three new adaptive smoothers which can choose optimal a parameter of the smoothers automatically, also we analyse them using a local Fourier analysis and our theorem to inform how to obtain an optimal parameter and the best smoother selection. Finally, various advantages of our recommended algorithm are illustrated, using both synthetic and real images.     

基于图像局部结构的区域匹配变分光流算法

针对变分光流算法的计算精度与鲁棒性问题,提出一种基于图像局部结构的区域匹配变分光流算法.光流估计能量泛函的数据项采用图像结构守恒与灰度守恒相结合,并引入规则化非平方惩罚函数,保证了光流估计的精度与鲁棒性;平滑项采用随图像局部结构自适应变化的扩散策略结合区域匹配约束函数能够有效地保护运动物体或场景的边缘轮廓信息;在光流计算过程中引入金字塔分层细化策略克服图像序列中大位移运动引起的像素点漂移现象,并采用数学方法证明光流估计模型的鲁棒性和收敛性.多组实验表明,本文方法在图像中存在剧烈光照变化、非刚性物体复杂运动以及多目标大位移运动等情况下具有较高的计算精度、较好的鲁棒性.     

Modeling of single-lap composite adhesive joints under mechanical and thermal loads

Two-dimensional (plane-stress and plane-strain) theoretical models are presented for stress analysis of adhesively bonded single-lap composite joints subjected to either thermal or mechanical loading or a combination thereof. The joints consist of similar/dissimilar orthotropic or isotropic adherends and an isotropic adhesive interlayer. The governing differential equation of the problem is obtained using a variational method which minimizes the complementary strain energy in the bonded assembly. In this formulation, through-thickness variation of shear and peel stresses in the interlayer is considered. Both shear and normal traction-free boundary conditions are exactly satisfied. Peel and shear stresses obtained from plane-strain analytical models considering a homogeneous adhesive interlayer are in close agreement with those of the finite element predictions. A systematic parametric study is also conducted to identify an ideal set of geometric and material parameters for the optimal design of single-lap composite joints.     

基于VMD-HHT边际谱的水工结构损伤诊断

为了准确识别水工结构的损伤,提出一种变分模态分解(variational mode decomposition,简称VMD)和Hilbert-Huang变换(Hilbert-Huang transform,简称HHT)边际谱相结合的水工结构损伤诊断方法。首先,采用联合的小波阈值和经验模态分解(empirical mode decomposition,简称EMD)降噪方法对原始信号进行降噪,减小环境噪声对结构损伤特征信息的干扰;其次,运用方差贡献率数据融合算法对降噪后各测点信号进行动态融合,提取结构完整的振动特性信息;然后,采用VMD方法将动态融合信号分解为一系列固态模量(intrinsic mode function,简称IMF),对各IMF分量进行Hilbert变换,求出融合信号的边际谱;最后,在VMD边际谱的基础上提取一种新的损伤特征向量-损伤灵敏指数,将其与马氏距离相结合对水工结构的损伤类型进行分类,并将该方法应用于悬臂梁模型试验。结果表明:该方法能够有效提取水工结构的损伤特性,准确识别水工结构的损伤和运行状态,为水工结构的安全运行提供了基础。     

全变分自适应图像去噪模型

通过分析三种主要变分去噪模型(调和、全变分以及广义全变分模型)的优缺点,提出了一种基于全变分的自适应图像去噪模型。该模型根据噪声图像的信噪比,采用高斯滤波器对图像进行预处理,克服了全变分模型引入的阶梯效应;利用图像中每一像素点的梯度信息,自适应选取去噪模型中决定扩散强弱的参数p(x,y),使接近边缘处平滑较弱,远离边缘处平滑较强。数值实验表明,本方法在去除噪声的同时保留了图像的细节信息,取得了很好的降噪性能,其峰值信噪比(PSNR)在高噪声水平下,较其他变分方法至少提高1.0dB左右。     

基于卷积变分自编码和多头自注意力机制的断路器剩余机械寿命预测

针对万能式断路器退化过程的不确定性,考虑到振动信号对机械性能退化的完善表征,提出了一种基于卷积变分自编 码(CVAE)和多头自注意力机制(MSA)的断路器分闸机械机构寿命预测方法。 首先依据断路器不同的事件区间提取参数特 征,再通过 CVAE 挖掘信号成分中的深度特征,将参数特征与深度特征融合得到完备退化特征,最后建立 GRU-MSA 的定量寿 命预测模型,引入了多头自注意力机制,在多个不同表征子空间中捕捉信号的不同依赖关系,对重要的时间步赋予更大的权重。 最后利用 3 台试品的振动信号测量数据对所提断路器分闸机械机构寿命预测方法进行测试,结果表明,所提出的方法在 3 个数 据集中寿命预测均方根误差(RMSE)分别为 141. 46、128. 75 和 134. 16,平均绝对误差(MAE)分别为 112. 17、101. 52 和 106. 22, 预测精度高且稳定性好,相对于其他混合预测模型更具优势。