然而由于聚类算法本身的局限性，重构信息无法完全反应。Li等人将经验模态

However, due to the limitations of clustering algorithm itself, information can not be reconstructed completely reacted. Li et al., The empirical mode decomposition algorithm (Empirical Mode Decomposition, EMD) introducing OLTC vibration analysis, and frequency domain characteristics diagnosed by mechanical failure load tap-tap, but the EMD decomposition algorithm nonlinear signal, will die aliasing state, cause the computing results are not very accurate

然而由于聚类算法本身的局限性，重构信息无法完全反应。Li等人将经验模态分解算法(Empirical Mode Decomposition, EMD)引入OLTC振动分析，并通过频域特性诊断出有载分接开关的机械故障，但是EMD算法在分解非线性信号时，会出现模态混叠现象，导致计算结果不是十分准确

However, due to the limitations of clustering algorithm, the reconstruction information can not fully reflect. Li et al. Introduced the empirical mode decomposition (EMD) into the OLTC vibration analysis, and diagnosed the mechanical fault of the on load tap changer through the frequency-domain characteristics. However, when the EMD algorithm decomposes the nonlinear signal, it will appear the phenomenon of mode aliasing, which leads to the calculation result is not very accurate