基于EEMD的交联电缆局部放电信号检测研究

摘要

针对经验模态分解方法在提取局部放电信号时易产生模态混叠问题，将集合经验模态分解方法用于交联电缆局部放电信号的提取。该方法相对传统小波变换不需要选择基函数和确定分解层数，可以地自适应地将含有干扰信号的局部放电信号分解成不同时间尺度的固有模态函数，对主要成分是局放信号的分量进行阈值处理与重构得到局放信号，经过希尔伯特-黄变换得到局放信号的时间-相位-幅值谱，获得反映局放信号的特征信息。仿真和实验结果表明验证了集合平均经验模态分解在交联电缆局部放电信号检测中的可行性。%To overcome the problem of mode mixing caused by EMD (Empirical Mode Decomposition)in extracting the signal of PD(partial discharge), EEMD(Ensemble Empirical Mode Decomposition) is adapted in extracting the XLPE cable PD signal. Compared with the traditional method wavelet transform, the method does not need to select basis function and fix decomposition level, the signal of PD with interference signal is decomposed into various IMFs(Intrinsic Mode Function) in time scales adaptively, then it limits the IMFs and reconstructs the IMFs to obtain PD signal, after operating by HHT(Hilbert-Huang Transform), the time-phase-amplitude spectrum which reflects the information of PD signal is obtained, the characteristics of the PD signal information is required. The results of simulation and experimental verify the feasibility of EEMD in detecting PD in XLPE cable.