基于GS-SVR和窗函数的EEMD端点效应抑制方法研究  

Research on EEMD Endpoint Effect Suppression Method Based on GS-SVR and Window Function

作　　者：李大帅 谷晓娇 LI Dashai;GU Xiaojiao(School of Mechanical Engineering,Shenyang University of Technology,Shenyang 110159,China)

机构地区：[1]沈阳理工大学机械工程学院,沈阳110159

出　　处：《机械工程师》2024年第10期41-43,48,共4页Mechanical Engineer

摘　　要：为了解决集合经验模态分解(Ensemble Empirical Mode Decomposition,EEMD)在分解过程中会出现端点效应的情况,文中提出了一种基于GS-SVR和窗函数的EEMD端点效应抑制方法,该方法首先采用网格搜索算法(Grid Search,GS)优化支持向量机回归(Support Vactor Regression,SVR)方法进行端点延拓,将延拓后的信号进行加窗处理,最后对加窗处理后的信号进行EEMD分解,进而得到有效的本征模态分量。最后基于振动仿真信号和实际采样信号,与传统EEMD分析方法比较以证明该方法的有效性。In order to solve the problem of endpoint effects during the decomposition process of Ensemble Empirical Mode Decomposition(EEMD),this paper proposes an EEMD endpoint effect suppression method based on GS-SVR and window functions.This method first uses the Grid Search(GS)algorithm to optimize Support Vector Regression(SVR)method for endpoint continuation.The extended signal is windowed,and finally the windowed signal is subjected to EEMD decomposition to obtain effective intrinsic mode components.Finally,based on vibration simulation signals and actual sampled signals,the effectiveness of the proposed method is verified by comparing it with traditional EEMD analysis methods.

关 键 词：网格搜索 支持向量机 EEMD 

分 类 号：TH133.33[机械工程—机械制造及自动化]