基于KPCA和WPHM的滚动轴承可靠性评估与寿命预测

为了评估滚动轴承的可靠性和预测剩余使用寿命,选取能够反映性能退化过程的特征参数作为寿命预测模型的输入参数,提出一种基于核主元分析(kernel principal component analysis,简称KPCA)和威布尔比例故障率模型(Weibull proportional hazards model,简称WPHM)的方法。首先,提取滚动轴承全寿命周期的时域、频域及时频域等多特征参数,从中筛选出有效的特征参数,构建高维相对特征集;其次,进行核主元分析,选取能够反映轴承全寿命周期性能退化过程的核主元,进而作为WPHM的协变量来进行可靠性评估和剩余寿命预测。通过滚动轴承全寿命试验,验证了该方法能够对轴承进行准确的可靠性评估和剩余寿命预测,以提供及时的维修决策。同时,由于提取的是相对特征,降低了同种轴承间在制造、安装及工况的差异,增强了该方法的适用性和稳定性。     

Deep Spatiotemporal Convolutional-Neural-Network-Based Remaining Useful Life Estimation of Bearings

The remaining useful life(RUL) estimation of bearings is critical for ensuring the reliability of mechanical systems. Owing to the rapid development of deep learning methods, a multitude of data-driven RUL estimation approaches have been proposed recently. However, the following problems remain in existing methods: 1) Most network models use raw data or statistical features as input, which renders it di cult to extract complex fault-related information hidden in signals; 2) for current observations, the dependence between current states is emphasized, but their complex dependence on previous states is often disregarded; 3) the output of neural networks is directly used as the estimated RUL in most studies, resulting in extremely volatile prediction results that lack robustness. Hence, a novel prognostics approach is proposed based on a time–frequency representation(TFR) subsequence, three-dimensional convolutional neural network(3 DCNN), and Gaussian process regression(GPR). The approach primarily comprises two aspects: construction of a health indicator(HI) using the TFR-subsequence–3 DCNN model, and RUL estimation based on the GPR model. The raw signals of the bearings are converted into TFR-subsequences by continuous wavelet transform and a dislocated overlapping strategy. Subsequently, the 3 DCNN is applied to extract the hidden spatiotemporal features from the TFR-subsequences and construct HIs. Finally, the RUL of the bearings is estimated using the GPR model, which can also define the probability distribution of the potential function and prediction confidence. Experiments on the PRONOSTIA platform demonstrate the superiority of the proposed TFR-subsequence–3 DCNN–GPR approach. The use of degradation-related spatiotemporal features in signals is proposed herein to achieve a highly accurate bearing RUL prediction with uncertainty quantification.     

圆角方形断面气动特性试验

通过刚性测压模型风洞试验,在均匀流场中测试了圆角率R/D=0.4的方形断面在0°～45°风向角、雷诺数Re=0.8×10~5～3.8×10~5时的气动特性。结果表明:试验雷诺数范围内,所有风向角下模型的阻力系数均值、升力系数均值、升力系数脉动值及斯特罗哈数基本上都受到雷诺数的影响。与其他风向角相比,0°～17.5°风向角时,阻力系数均值和升力系数均值受雷诺数的影响更大; 7.5°～45°风向角时,升力系数脉动值受雷诺数的影响更大; 0°～12.5°风向角时,斯特罗哈数受雷诺数的影响更大。     

地铁环境振动源强测试与评价标准分析

实测南昌地铁1号线隧道内约200趟列车通过邻近非减振与减振断面的振动响应。基于不同振动评价标准,从统计角度分析了不同车次列车运行对时域和频域的影响、测点位于隧道壁不同高度对振动源强值及钢弹簧浮置板减振量评价的影响。分析结果表明:钢弹簧浮置板隧道壁上的减振效果明显,但浮置板本身的振动响应大幅增加。不同车次对各测点低频段及隧道壁高测点影响较大,但对Z振级影响较小。振动在隧道壁上低频段有所放大。由隧道壁低测点测得的最大Z振级最适合评价地铁源强值,南昌地铁实测源强值为76.66 d B。不同高度测点及不同评价标准对浮置板减振量评价有较大影响,建议采用低测点最大Z振级评价浮置板减振量。     

直升机主减半主动反共振隔振技术研究

针对滚动轴承剩余寿命预测中的特征评估及模型优化问题,提出了面向轴承寿命的特征评估与模型优化的方法。该方法在轴承特征进行单调性与敏感性评估的基础上,对轴承运行状态跟踪能力进行量化评估,进而筛选出表征轴承性能退化的多维特征集。为了减少多维特征集之间相关冗余信息对寿命预测的影响,采用相似近邻传播（affinity propagation ,简称AP）聚类方法对多维特征集进行聚类和筛选。为了统一描述AP聚类后的特征对轴承退化状态的表征信息,采用自组织神经网络（self-organizing feature map ,简称SOM）构建轴承健康指数。最后,利用自适应混沌粒子群算法（adaptive chaos particle swarm optimization, 简称ACPSO）优化双指数模型预测轴承剩余寿命。试验表明,该方法可以准确描述轴承运行状态时期,并有效地预测了轴承的剩余寿命。     

气动准零非线性隔振器的刚度特性与参数调控

针对大载荷、低动态频率的机械装备的隔振需求,提出了气动准零刚度非线性隔振技术方案,利用气体负压为静载、波纹管结构提供动刚度,实现了隔振装置的高静态低动态刚度特性。阐述了光滑型气动准零隔振器的结构及工作原理,根据气体状态方程推导了隔振装置的非线性刚度理论模型。为了得到符合设计要求的刚度特性,分别研究了气体占比与初始压强等参数对隔振装置刚度特性的调控规律,发现了合理的气/液比例,既可以实现较低的动态刚度,同时起到大幅振动下位移限位的效果。探讨了过载加速度对隔振系统固有特性的影响,研究发现,过载的存在使得隔振静平衡位置偏离设计的平衡点,因此其线化系统固有频率会发生较明显的变化,这是工程设计根据实际情况需要考量的一点。     

基于递归熵及长短期记忆神经网络的滚动轴承退化趋势预测

针对滚动轴承退化趋势预测中的特征选择及模型优化问题,提出了基于相空间重构结合长短期记忆神经网络的滚动轴承退化趋势预测方法.首先,在滚动轴承退化特征灰色关联分析的基础上,对滚动轴承运行状态跟踪能力进行量化评估;然后,为统一描述各个特征对轴承退化状态的表征信息,采用谱聚类方法对滚动轴承运行状态进行了准确划分;最后,以滚动轴...     

气动准零非线性隔振器的刚度特性与参数调控

为解决滚动轴承剩余寿命预测问题,提出一种基于长短期记忆网络（long short-term memory, 简称LSTM）的剩余寿命预测方法。首先,从时域、频域及时频域特征中提取特征参数;其次,定义三个评价指标定量评估表征轴承退化过程的特征参数效果,筛选得到退化特征参数集,搭建长短期记忆网络预测模型并以归一化寿命值为标签训练神经网络;最后,用训练好的神经网络实现滚动轴承剩余寿命预测。通过滚动轴承全寿命试验证明,该方法可以准确预测滚动轴承剩余寿命,并与反向传播(back propagation machine,简称BP)神经网络和支持向量回归机（support vector regression machine,简称SVRM）的预测效果对比,验证了提出方法的有效性。     

滚动轴承剩余使用寿命预测综述

滚动轴承作为旋转机械的关键零部件,其剩余使用寿命(RUL)预测对生产维修和人身安全具有重要意义。由于滚动轴承复杂多变的工作环境,使得同工况的参考样本少而变工况的参考样本较多,具有不平衡、不完整、无标签及噪声干扰等特性,增加了滚动轴承RUL预测的困难。随着大数据时代的来临和人工智能的发展,滚动轴承RUL预测方法也变得更加丰富。因此,在故障预测与健康管理(PHM)的框架下,对滚动轴承失效模式和故障数据特点进行阐述,对故障特征提取、降维和融合方法以及得到的性能退化指标分别进行了分类和对比分析。结合数据驱动算法,对滚动轴承RUL的预测方法、模型选择和评估标准进行了梳理和对比。最后对滚动轴承RUL预测未来的发展趋势进行了展望。     

基于局部均值分解与拉普拉斯特征映射的滚动轴承故障诊断方法

针对滚动轴承非平稳振动信号的特征提取及维数优化问题,提出了融合局部均值分解与拉普拉斯特征映射的轴承故障诊断方法。首先,通过局部均值分解对非平稳振动信号进行平稳化分解,提取乘积函数分量、瞬时频率及瞬时幅值的高维信号特征集;然后,将高维特征集作为拉普拉斯特征映射算法的学习对象,提取轴承高维故障特征集的内在流形分布,以获得敏感、稳定的轴承振动特征参数,实现基于非平稳振动信号分析的滚动轴承故障特征提取;最后,结合支持向量分类模型量化LMD-LE方法的特征提取效果,实现不同状况下的轴承故障分类。轴承故障样本分类识别平均正确率达到91.17%,表明LMD-LE方法有效实现了高维局部均值分解特征集合的降噪,所提取的特征矩阵对轴承故障特征描述准确。     

一种融合多传感器数据的数模联动机械剩余寿命预测方法

随着传感和信息技术的发展,各式各样的传感器获取了机械装备海量的监测数据,让剩余寿命预测有"据"可依,推动机械剩余寿命预测进入了大数据时代。但由于数据类型多样、量大面广,如何利用丰富的多传感器数据,从中快速挖掘健康状态退化信息,指导寿命预测,成为大数据时代下机械寿命预测的全新挑战。基于模型的寿命预测方法大多仅针对单一监测数据进行建模分析,无法有效利用丰富的大数据资源。数据驱动的方法则过分依赖训练数据,缺乏必要的经验指引,方法的可解释性差。为了有效利用多传感器数据指导寿命预测,从数模联动的思路出发,建立了一种融合多传感器数据的数模联动寿命预测方法。采用一种通用的Wiener过程模型对健康状态退化过程进行描述,分别建立多源观测函数和多源映射函数对状态与数据之间的因果关系和关联关系进行描述,采用粒子滤波算法将多传感器数据与模型进行动态匹配,预测剩余寿命。在提出方法的统一框架指导下,选取三种特定模型对铣刀剩余寿命进行预测,验证了提出方法的有效性。     

基于Kriging函数的KVPMCD在滚动轴承故障诊断中的应用

滚动轴承的故障诊断本质上是模式识别的问题,多变量预测模型(VPMCD)是一种新的模式识别方法,其实质就是通过特征值之间的相互内在关系建立数学模型,并根据数学模型对被诊断轴承的特征值进行预测从而达到模式识别的目的。但是VPMCD分类方法中单纯采用回归模型进行预测,因此当故障特征值之间关系较为复杂时将导致预测精度降低。针对这一缺陷,提出了基于Kriging函数的多变量预测模型(KVPMCD)模式识别方法。Kriging模型由回归模型和相关模型组合而成,其中,相关模型是在全局模型的基础上创建的局部偏差,它恰恰可以揭示特征值之间的空间相关性,从而弥补原VPMCD中单纯采用回归模型的缺点。对UCI标准数据以及滚动轴承实测数据的分析结果表明,KVPMCD模式识别方法比原VPMCD方法可以更加有效地识别滚动轴承的工作状态和故障类型。     

Rolling element bearings absolute life prediction using modal analysis

In this paper, the authors introduce an experimental procedure for predicting the fatigue life of each individual rolling element bearing separately using vibration modal analysis. The experimental procedure was developed based on a statistical analysis. A statistical analysis was performed to find an empirical model that correlates the dynamic load capacity of rolling bearings to their dynamic characteristics (Natural frequencies and damping). These dynamic characteristics are obtained from the frequency response function of each individual bearing that results from vibration modal analysis. A modified formula to the already known Lundberg-Palmgren life formula is proposed for rolling element bearings. Given the modified formula, one can predict the fatigue life of each individual rolling element bearing based on its dynamic characteristics. The paper compares the results from the modified formula with those from Lundberg-Palmgren formula. The modified formula provides an accurate prediction for the fatigue life of each individual bearing based on its dynamic characteristics. The experimental validation of the modified formula is considered for future work. Therefore, it can be used in various applications of rolling element bearings in machinery systems.     

On the flow of emulsions in tribological contacts

There are numerous instances of technical importance in which multi-component lubricants are utilized either by design or by necessity. In many of these cases one of the components is a liquid while the other component is a gas, or both components are liquids. These mixtures do not exhibit Newtonian behavior even when their components themselves are Newtonian, and thus classical lubrication theory is inapplicable to bearings lubricated with mixtures. Our objective here is to extend hydrodynamic lubrication theory to lubrication with liquid-liquid and liquid-gas mixtures. The extended theory is then applied to journal bearings, roller bearings and to lubrication in cold rolling. We are able to predict several experimentally observed phenomena such as oil-pooling ahead of an elastohydrodynamic lubrication conjunction.     

Residual fatigue life prediction of ball bearings based on Paris law and RMS

Paris law can reflect the failure mechanism of materials and is usually used to be a method to predict fatigue life or residual fatigue life.But the variable which can represent the health of machine i...     

Particle Damage in Hertzian Contacts and Life Ratings of Rolling Bearings

Particle denting, and contamination marks found on bearing raceways, can induce stress concentrations and facilitate surface initiated fatigue. The lubricant film developed at the dent and related local surface stresses are also significant to the crack initiation mechanism. In this article, a new methodology is presented to link the micro-EHL film and related local stresses to the fatigue life of rolling bearings. The applied methodology is based on Fourier analysis of the harmonic components of the surface microgeometry to predict stresses and induced lubricant film. The application of this method to actual bearing surfaces is discussed and analyzed in relation to some existing microcontact EHL solutions. A global evaluation of the use of the method to rolling bearings dynamic load ratings is also carried out. A comparison between experimentally obtained rolling bearing life and lives predicted using the present theory indicate the global ability of the model to describe the effect of the lubrication quality on the life expectancy of rolling contacts. From this analysis, an assessment of some typical equations used in rolling bearing dynamic ratings is carried out. It is found that the degree of lubrication of the rolling contact and the cleanliness conditions of the oil are indeed significant to the prediction of the life expectancy of the bearing.     

基于状态空间模型的飞机APU在翼RUL预测方法

为解决利用飞机辅助动力装置(APU)在翼监测数据难以表征其性能状态而造成的性能评估以及剩余使用寿命预测(RUL)难的问题,本文提出一种基于状态空间模型(SSM)与卡尔曼滤波融合的APU在翼RUL预测方法.首先,通过在翼监测数据构造含噪声的性能指标(PI)来表征APU的性能状态,借助维纳过程与建立的含噪声的PI构建状态方...     

多传感器融合和 MHA-LSTM 的电机轴承剩余寿命预测

轴承作为电机的核心部件, 主要起到支撑引导轴、 减小设备摩擦、 连接不同设备等作用, 其剩余寿命预测对系统健康 管理起着十分重要的作用。 针对单一传感器信号通常难以全面描述系统的潜在退化机制, 论文提出一种基于多头注意力机制 和长短时记忆神经网络的电机轴承剩余寿命预测模型。 首先, 基于马氏距离确定轴承性能退化起始点, 将滚动轴承全寿命周 期分为正常阶段与退化阶段; 其次, 使用自编码器自动提取振动信号特征, 并将其与电机电流、 轴承温度融合, 构成多源信息 特征矩阵; 然后基于多头注意力机制和长短时记忆网络模型动态选择相关度较高的特征, 提高寿命预测的准确性。 最后, 采 用实验数据进行验证, 结果表明所提出的模型具有更高的准确性。     

基于谱峭度和AR模型的滚动轴承故障诊断

提出基于自回归(Autoregressive,简称AR)预测滤波的谱峭度分析方法,将其应用于滚动轴承的早期故障诊断。通过结合AR预测滤波器提取轴承故障信号共振衰减成分的特性,利用谱峭度方法对AR预测滤波器滤波后的信号进行处理,实现了滚动轴承早期微弱故障的识别。通过滚动轴承的疲劳全寿命加速实验获取滚动轴承的自然故障信号,克服了传统轴承故障诊断人工加工故障的不足。通过试验数据的分析表明,基于AR预测滤波的谱峭度方法不仅能够消除干扰成分提取故障特征,还能增加谱峭度方法的稳定性。     

[数字孪生驱动的智能制造]数字孪生与深度学习融合驱动的采煤机健康状态预测

针对处于恶劣工作环境的采煤机状态预测与维护困难的问题,结合数字孪生高逼真度行为仿真特性和深度学习强大的数据挖掘能力,提出数字孪生与深度学习融合驱动的采煤机健康状态预测方法。基于物理空间多物理参数构建采煤机数字孪生体,通过在虚拟空间的可视化展示与分析实现健康状态预判;建立基于深度学习的采煤机关键零件剩余寿命预测模型,实现实时监测数据驱动下的零件剩余寿命的在线预测;综合数字孪生体状态和剩余寿命值,实现采煤机健康状态预测。通过试验验证了该方法的有效性,为采煤机健康状态预测与管理提供新思路。