基于镜像延拓和神经网络的EMD端点效应改进方法

经验模态分解(Empirical Mode Decomposition,EMD)在机械故障诊断中存在一个比较严重的问题,即端点效应。镜像延拓是克服端点效应的有效方法,采用镜像延拓处理端点效应时,要求将镜面放置在局部极值点处,针对这一问题,提出了一种基于镜像延拓和神经网络相结合的数据延拓方法。采用神经网络预测原信号序列,将信号向前向后各延拓一个极值点,再采用镜像延拓有效地减小EMD分解中的端点效应。通过对仿真信号的分析,验证了该方法能有效抑制EMD方法中的端点效应问题。

Improved Method for End Effects of EMD Based on Mirror Extension and Neural Network

Empirical Mode Decomposition (EMD) has been employed extensively in mechanical fault diagnosis,but the serious problem existed in the application is end effects. Mirror Extension is an effective method to overcome it. When the mirror extension is used to treat with end effects,it needs to put the mirror at local extreme point. Aiming at this problem,a method combining the mirror extension with the Neural Network is proposed to extend the data. The Neural Network is adopted to forecast an extreme point forwa...