说明：  提出了将信号进行相空间重构后再采用奇异值分解, 对分解后的主成分进行包络分析, 从而提取信号的隐含特 征的方法, 并将该方法应用于齿轮的局部故障振动特征信号的提取中。数值仿真实验结果表明, 该方法能有效提取强背景 信号及噪声中的弱冲击特征信号, 是一种有效的弱信号特征提取方法。采用该方法对齿轮振动信号进行故障特征提取与识 别, 结果与实际情况相符。
(Signal implicit characteristic of phase space reconstruction, and then using the singular value decomposition (SVD), principal component decomposition envelopment analysis, so as to extract the signal, and the method is applied to the partial failure of the vibration characteristics of the signal of the gear extraction. The numerical simulation results show that this method can effectively extract Weak Feature strong background signal and noise in the signal, a weak signal feature extraction methods. The extraction and recognition of fault feature of gear vibration signal results consistent with the actual situation.)