由于排列熵算法能够有效放大时间序列的微弱变化，且计算简单、实时性好，已在信号突变检测方面显示出良好的应用前景，但是排列熵算法中嵌入维数和延迟时间等参数的确定仍依赖于经验和尝试，该问题已成为排列熵算法走向工程应用的瓶颈问题。根据排列熵算法的原理，提出了基于重构时间序列最佳相空间来确定模型参数的方法。根据相空间重构的两种观点，介绍了延迟时间与嵌入维数独立确定和联合确定两种方法的基本理论，然后利用仿真信号和滚动轴承全寿命数据对两种算法进行了检验和对比。结果表明，模型参数的独立确定方法比联合确定方法对信号的异常检测更好。%Permutation entropy (PE)algorithm can better magnify tiny change of a time series of data.It is simple in computation and shows good quality in real-time application,so,it gives us a good application prospect in detection of the sudden change of a signal.However,the parameters in the algorithm,namely the embedding dimension and delay time are usually still determined by experience or trial.This forms a bottle-neck of PE algorithm for engineering application.According to the theory of PE algorithm,a method based on reconstructing optimal phase space of time series was put forward to determine these model parameters.Considering two points of view about phase space reconstruction, basic theories of independent and joint determination methods were introduced to determine the delay time and embedding dimension.The two determination methods were validated and compared by using simulated signals and whole life data of rolling bearings. It is concluded that the independent determination of model parameters was better than joint determination for abnormality detection.
展开▼
