排列熵算法研究及其在振动信号突变检测中的应用

Research and application of the arithmetic of PE in testing the sudden change of vibration signal

排列熵算法具有计算简单、实时性高、能较好地反映时间序列数据微小的变化,已成为一种检测复杂系统动力学突变的有力工具。首先阐述了排列熵算法的基本原理,通过仿真信号的排列熵计算结果,说明了时间序列排列熵对信号突变的检测效果,最后以某型坦克变速箱振动信号为例,通过提取振动信号的峰峰值特征,形成新的时间序列数据用于排列熵的计算,结果表明排列熵值的变化能够有效反映信号的突变时刻。

Since permutation entropy（PE） algorithm can be computed simply,shows good quality in real-time application and can better distinguish tiny change of a time series data,it has become a powerful tool in detecting sudden changes of a complex dynamic system.Basic principle of PE algorithm is introduced in this paper firstly,and then a simulated signal is taken as an example to verify the effectiveness of PE value in abnormal detection.Finally,vibration signal from a gearbox of a certain tank is analyzed,features of peak-to-peak value are calculated to form a new series data.PE algorithm is applied to the new peak-to-peak series data.Results show that PE can find the moment corresponding to sudden change of the vibration signal successfully.