基于扩展双谱的齿轮压缩采样信号特征提取方法

Feature Extraction Method of Gear Compression Sampling Signals based on Extended Bispectrum

针对传统信号处理方法不适用于压缩采样信号的问题,提出一种基于扩展双谱的压缩信号特征提取方法,以实现对压缩后的齿轮振动信号直接进行特征提取而避免复杂的信号重构过程。该方法首先将原始信号与M序列进行混频,将高频信息搬移到低频部分,低通滤波后通过低频采样获取压缩采样信号,再对压缩信号进行扩展双谱分析,针对实际应用中存在的问题,进一步引入阶次分析,角域同步平均等技术。仿真及工程实验表明,该方法能有效地提取出压缩采样信号的相位耦合特征,实现故障识别。

Traditional signal processing method is not suitable for compressive sampling signals.In this paper,a feature extraction method of compressive signals based on extended bispectrum is proposed to realize the direct feature extraction of compressive gear vibration signals and avoid the complicated process of signal reconstruction.Firstly,the original signal is mixed with the M sequence,and the high frequency information is shifted to the low frequency part.After low-pass filtering,the compressive sampling signal is acquired by low-frequency sampling.Then,the compressive signal is analyzed by the extended bispectral analysis.Techniques of order analysis and angular domain synchronous averaging are used to solve the problems in practical applications.Simulations and engineering tests show that this method can effectively extract the phase coupling characteristics of compressive sampling signals and realize fault identification.