中文摘要：

介绍了经验模态分解故障诊断方法，该方法统一了瞬时频率的概念，产生时频域分析方法一本征模态函数，可以突出局部数据特征，提取更准确、更有效的原始信号特征信息，并经过分解，提炼出有效时域信号，对其进行Hilbert-Huang变换，实现信号在频域中的再分析：提出的激光多普勒测振技术，具有抗干扰、高分辨率、高精度、非接触式的振动优点。通过激光测振仪采集数控机床齿轮振动信号，并借助经验模态分解方法，诊断出轴承与轴瓦之间存在着频率为33-3Hz的周期性摩擦现象，从而证明了EMD能从大量的非线性、非平稳信号中提取振动特征信息和相关的模态参数，是一种非线性、非平稳等时变信号处理方法。

英文摘要：

Experience mode decomposition fault diagnosis method is introduced, the method unifies the concept of instantaneous frequency and promethean combination of time domain and frequency domain, generates the time-frequency domain analysis method, the intrinsic mode function, can highlight local characteristics of data, extract more accurate, more effective original signal characteristic information, and through the decomposition, extract the time domain signal effectively, through the Hilbert-Huang transform, to realize signal＇s reanalysis in the frequency domain; Laser Doppler vibration technology, has advantages such as anti-jamming, high resolution, high accuracy, non-contact. The subject collects the gear vibration signals ofnc machine tool through laser vibrometer, and using empirical mode decomposition method, diagnoses that there exists the periodic friction phenomenon at 33.3 Hz frequency between the bearing and bearing shell, which proves that the EMD can extract vibration characteristic information and the relevant modal parameters from a large number of nonlinear and non-stationary signals, and it is an excellent method in processing nonlinear, non-stationary and time-varying signals.