摘要
对于供输弹系统早期故障信号非线性、非平稳、故障信息易被湮没难以识别的问题,提出了基于优化k值的变分模态分解(VMD)和矩阵分形的早期故障识别方法。首先计算不同k值下相对能量比值μ的大小,选择与较小μ值对应的k作为分解层数并对分解结果进行检验;然后通过计算VMD分解后得到的各工况信号IMF的广义维数,构建每种工况的分形矩阵;最后计算样本信号与待测信号分形矩阵的相关系数,确定待测信号的工况。结果表明该方法能对供输弹系统早期故障进行有效识别。
For the problems of nonlinear and nonstationary fault signals in the early stage of supply and delivery system,the fault information is easy to be buried and difficult to identify,the early fault diagnosis methods of k value Optimization for Variational Mode Decomposition(VMD)and matrix fractal are put forward.Firstly,the relative energy ratioμof different k values was calculated,and the decomposition layer number ofμwas selected to test the decomposition results,then the fractal matrix of each condition is formed by calculating the broad dimension of the signal component of each condition obtained by the VMD decomposition,and the correlation coefficient between the sample signal and the fractal matrix of the signal to be measured is calculated to determine the condition of the signal to be measured.The results show that the method can effectively identify the early fault of the bomb-supply system.
作者
席茂松
许昕
潘宏侠
XI Maosong;XU Xin;PAN Hongxia(School of Mechanical Engineering,North University of China,Taiyuan 030051,China;System Identification and Diagnosis Technology Research Institute,North University of China,Taiyuan 030024,China)
出处
《机械设计与研究》
CSCD
北大核心
2020年第2期208-211,215,共5页
Machine Design And Research
基金
国家自然科学基金资助项目(51675491)
面上自然基金项目(201801D121185)。
关键词
优化k值
变分模态分解
分形矩阵
广义维数
k value optimization
variational modedecomposition
fractalmatrix
the generalized dimension
