线性外推法和相似极值法耦合抑制EMD端点效应
The Coupling of Linear Extrapolation Method and Similar Extreme Extension Method to Restrain Boundary Effect of EMD

在应用经验模态分解（Empirical mode decomposition，EMD）进行机械故障诊断过程中，存在上下包络线拟合时产生端点效应的现象。在线性外推法和相似极值法的基础上，考虑信号内部的极值变化稳定性的情况，对EMD端点延拓进行了改进。通过计算信号内部的方差，确定信号内部稳定性的强弱。在信号内部变化平稳但稳定性较弱的情况下，采用相似极值延拓法，最大程度地保护信号内部的平稳性；在信号内部波动性较强的情况下，考虑到信号内部发生异常，采用线性外推法对信号两端进行预测延拓。模拟信号仿真和滚动轴承滚动体故障的实例分析。结果表明，该方法能有效地抑制EMD方法的端点效应。
The end effect exists in envelope fitting during the process of applying EMD method to diagnose mechanical fault. A revised method of boundary extension is proposed based on linear extrapolation method and similar extreme extension method considering the internal signal extreme waving condition in this paper. The variance of internal signal is calculated to evaluate the degree of signal stability. In the case of the signal waves smoothly, similar extreme extension method is conducted to maintain the inherent steadiness at utmost. While the signal that changes abnormally swings violently, the local information at the edges is considered only to conduct linear extrapolation method. Through simulation analysis and tests of fault diagnosis of rolling bearings, results demonstrate this method can suppress end effect of EMD method