经验模态分解在切削振动信号分析中的应用

Application of empirical mode decomposition in analyzing cutting vibration signal

为了分析切削振动的变化规律及产生根源,采用在线监测的方法采集振动信号,提出了运用经验模态分解对振动信号进行分析的方法,构建了分段频率变化的仿真振动信号,分别对仿真振动信号进行短时傅里叶变换、小波变换和经验模态分解时频分析。结果表明,经验模态分解方法具有更高的时频分辨率,运用经验模态分解方法对实际振动信号进行分析,提取典型分量绘制希尔伯特幅值谱,能较好地反映振动信号的时频变化规律。经验模态分解适合对切削振动信号进行分析,但是需要对算法的模态混叠及计算效率低等不足进行完善。研究结果对非平稳信号的分析具有参考价值。

In order to analyze the change law and the origin of the cutting vibration,the on-line monitoring method is used to collect the vibration signal.A method of analyzing vibration signals by using empirical mode decomposition（EMD）is proposed.A simulated vibration signal of piecewise frequency change is constructed.The short-time Fourier transform,wavelet transform and EMD are carried out respectively to the simulation vibration signal.The results show that the EMD method has higher time-frequency resolution.Moreover,when EMD method is used to analyze actual vibration signal and extract a typical component for drawing Hilbert amplitude spectrum,it shows that the EMD method can better reflect the time-frequency variation of the vibration signal.The EMD method is suitable for the analysis of cutting vibration signal.However,it is necessary to improve the mode mixing and low computational efficiency of the algorithm.The research result has some reference value for the analysis of non-stationary signals.