基于主分量分析的柴油机振动信号特征提取

针对柴油机振动信号非线性非平稳性的特点，提出一种相空间重构理论、局域波法与主分量分析相结合的信号特征提取模型，该模型首先应用相空间重构理论从已知时间序列中抽取动力系统，然后通过主分量提取以降低空间维数、突出故障信息，最后使用局域波时频分析方法对提取的主分量进行分析。通过对6BB1型柴油机实测信号进行的特征提取与分析表明，该方法能去除柴油机振动信号局域波时频图中的冗余信息，突出故障信息，从而证明了方法的有效性。

Feature Extraction from Vibration Signals of Diesel Based on PCA

The Local Wave method was combined with Principal Component Analysis(PCA) and Reconstruction theory as a model of feature extraction.In this model,reconstruction theory was used to extract dynamic space from time series,PCA was applied to reduce the dimension of the space and made the fault information clearly,and local wave time-frequency analysis method was used to recognize the faults.In the end,an example of practical application shows that the fault information of the vibration signals of 6BB1 Diesel engine is maked clearly by using the model above,the practicality is explained in reason.The example proves that this integrated method is feasible.