基于高斯混合模型的物流非高斯随机振动损伤分析  

Non-Gaussian random vibration damage analysis of logistics based on a Gaussian mixture model method

作　　者：郭涛 葛长风 夏斯璇 殷诚 林康 钱静 GUO Tao;GE Changfeng;XIA Sixuan;YIN Cheng;LIN Kang;QIAN Jing(School of Mechanical Engineering,Jiangnan University,Wuxi 214122,China;Department of Packaging Science,Rochester Institute of Technology,New York 14623,USA;Jiangsu Key Laboratory of Advanced Food Manufacturing Equipment&Technology,Wuxi 214122,China)

机构地区：[1]江南大学机械工程学院,江苏无锡214122 [2]罗切斯特理工学院包装科学系,美国纽约14623 [3]江苏省食品先进制造装备技术重点实验室,江苏无锡214122

出　　处：《振动与冲击》2024年第12期203-211,共9页Journal of Vibration and Shock

基　　金：国家重点研发计划(2018YFC1603300)。

摘　　要：针对公路运输环境中的振动信号具有明显的非高斯性,提出一种非高斯随机振动疲劳损伤分析方法。为了描述振动信号的幅值概率密度分布,采用移动加速度均方根来代表该段信号的振动强度,并引入高斯混合模型对加速度均方根值进行描述。在此基础上结合Tovo-Benasciutti方法和Dirlik方法推导出非高斯宽带频域疲劳损伤计算方法。最后,以雨流计数法作为参考,对不同峭度的实测振动信号进行疲劳损伤分析,结果表明,与传统频域疲劳损伤计算方法相比较,提出的非高斯疲劳损伤方法具有更高的计算精度。该研究对于运输包装件的随机振动加速试验设计有实际意义。Aiming at the obvious non-Gaussian property of vibration signals in transportation environment,a non-Gaussian random vibration damage analysis method based on a Gaussian mixture model was proposed.To describe the amplitude probability density distribution of the vibration signal,the moving root mean square of acceleration was introduced to represent the vibration intensity of the signal,and the Gaussian mixture model was used to describe the probability density distribution of root mean square of acceleration.On this basis,combined with the Tovo-Benasciutti method and the Dirlik method,a non-Gauss broadband frequency domain fatigue damage calculation method was derived.Finally,the fatigue damage analysis of measured vibration signals with different kurtosis was carried out with rain flow counting method as a reference.The results show that compared with traditional fatigue damage calculation methods in frequency domain,the calculation accuracy of the proposed non-Gaussian fatigue damage method was significantly improved.

关 键 词：非高斯随机振动 高斯混合模型 概率密度函数 运输包装 

分 类 号：TH212[机械工程—机械制造及自动化] TH213.3