基于DFT的布林线视角下的信号异常值检测方法

针对装备健康管理中关键部件常出现状态异常、性能退化问题,提出了1种兼顾时域统计性和频域周期性的信号数据异常值检测方法.在频域上,通过离散傅里叶变换(Discrete Fourier Transform,DFT)进行频谱分析,根据幅值大的低频谐波的周期选择布林线的计算周期;在时域上,根据不同的置信度,利用移动平均值和标准...

The Outlier Detection Method on Signal from the Perspective of Bollinger Bands based on DFT

Aiming at the problem of abnormal state and performance degradation of key components in equipment healthmanagement, an abnormal signal data detection method that takes into account both time domain statistics and frequencydomain periodicity is proposed. In the frequency domain, the discrete Fourier transform is used to analyze the frequencyspectrum, and the calculation period of Bollinger bands is selected according to the period of low frequency harmonicswith large amplitude. In the time domain, according to different confidence levels, tracks of Bollinger bands are deter?mined by the moving average and standard deviation, and abnormal areas are divided into different degrees for the operat?ing status of key components of the equipment. The analysis of positive and negative calculation examples shows that themethod is correct and effective, and provides theoretical guidance for missile equipment maintenance decision-making.