完全互补小波噪声辅助集总经验模态分解.pdf

振 动 与 冲 击 第 36卷第4 期 JOURNAL OF VIBRATION AND SHOCK Vol.36 No. 4 2017 完全互补小波噪声辅助集总经验模态分解 何 刘 ，丁建明，林建辉，刘新厂 (西南交通大学牵引动力国家重点实验室，成 都 6 10031) 主 商 要 ：经验模态分解(EMD )是一种自适应非线性非平稳数据处理方法。噪声辅助的EM D 方法能克服EM D 方 法在处理间歇信号时出现的“模态混叠”现象。在这些噪声辅助方法中，互补集总经验模态分解（CEEMD )和完全噪声辅 助噪声集总经验模态分解（CEEMDAN )恢复了 EM D 分解的完整性。在现有分析方法上提出了完全互补小波噪声辅助集 总经验模态分解（CCWEEMDAN )算法。该算法能用更小的集总数、更少的迭代次数和极小的计算消耗获得更好的光谱 分离效果和数目较少的筛选模态。 关键词：经验模态分解;集合经验模态分解;噪声辅助;模态混叠;互补集总经验模态分解 中图分类号：U 211;U 270 文献标志码： A D 0 I 10. ： 1 3 4 6 5 /j. cnki. jvs. 2 0 1 7 .0 4 .0 3 7 A complete complementary wavelet ensemble empirical mode decomposition with adaptive noise HE Liu， DING Jianming， LIN Jianhui， LIU Xinchang ( S tate Key Laboratory of Traction Power, Southwest Jiaotong U niversity, Chengdu 6 10 03 1 , China) Abstract ： E m p iric a l mode decom position (E M D ) is a self-adaptive m ethod and suitable to analysing the non ­ stationary and nonlinear signals . Noise-assisted versions have been proposed to alleviate the so-called “ mode m ix in g ” phenom enon , w hich may appear when an E M D algorithm is used to deal w ith a signal w ith in te rm itte n cy . Am ong th e m , the com plem entary ensem ble E M D (C E E M D ) and com plete ensemble E M D w ith adaptive noise (C E E M D A N ) recover the completeness property o f E M D . In this w ork a new algorithm named com plete com plem entary wavelet ensem ble em p irica l mode decom position w ith adaptive n o is e ( C C W E E M D A N ) was presented based on those existing tech niq ue s , obtaining better spectral separation o f the modes w ith few er siftin g iter