基于VMD的旋转机械故障诊断方法研究

发布时间：2018-07-04 12:29

本文选题：VMD + SVD　；参考：《石油矿场机械》2016年08期

【摘要】：旋转机械结构复杂,振动信号信噪比低且多为非平稳、非线性的多分量信号,出现故障时难以有效地进行诊断。常规的小波分析方法需根据信号特点选取特定的小波基和分解层次,自适应分解方法如EMD、EEMD等存在频率混叠及虚假分量现象,在提取微弱信号时易造成误判。提出了一种基于变分模态分解(VMD)和奇异值分解(SVD)的故障诊断方法。首先对信号进行VMD分解,并对分解得到的固有模态函数分量进行SVD降噪;然后从降噪后的分量中选取故障特征分量进行时频域及包络谱分析,最终确定故障类型。仿真及试验结果表明,该方法可以有效地降低噪声,提取微弱故障信息,实现故障诊断。
[Abstract]:Rotating machinery has complex structure, low signal-to-noise ratio (SNR) of vibration signals and non-stationary, nonlinear multi-component signals, so it is difficult to diagnose effectively when faults occur. Conventional wavelet analysis methods need to select specific wavelet bases and decomposition levels according to the characteristics of signals. Adaptive decomposition methods such as EMD-EEMD have frequency aliasing and false component phenomena which can easily lead to misjudgment when extracting weak signals. A fault diagnosis method based on variational mode decomposition (VMD) and singular value decomposition (SVD) is proposed. Firstly, the signal is decomposed by VMD, and the natural mode function component is decomposed by SVD, and then the fault characteristic component is selected from the noise reduction component to analyze the time-frequency domain and the envelope spectrum, and finally the fault type is determined. Simulation and experimental results show that this method can effectively reduce noise, extract weak fault information and realize fault diagnosis.
【作者单位】：中国石油大学(北京)机械与储运工程学院;中国石油新疆油田分公司实验检测研究院;中国石油西南管道公司贵阳输油气分公司;
【基金】：国家自然科学基金资助(51504274)
【分类号】：TH17

【相似文献】