爆破监测信号多为含噪信号，噪声会使经验模态分解（EMD）的结果产生严 重的模态混淆，使用改进算法 EEMD 对模态混淆有一定的抑制作用但效果并不明显。 为此本研究将使用自适应补充集合经验模态分解（CEEMDAN）来处理含噪信号。比较 EMD、EEMD、 CEEMDAN对仿真信号的分解结果，计算 EMD、EEMD、 CEEMDAN 得 到的 IMF 的排列熵值，对 EMD、EEMD、 CEEMDAN 的分解结果进行 Hilbert 变换，并比 较三者时频谱的分辨率。最后将 CEEMDAN 用于水下钻孔爆破地震波时频分析中，结果 表明： CEEMDAN 不仅对模态混淆具有一定的抑制作用，且其分解结果经过 Hilbert 变 换得到的时频谱在时域和频域上都具有较高的分辨率。

Blasting monitoring signals are mostly noisy signals, and nosie can cause serious mode confusion in empirical mode decomposition (EMD) results. The improved EEMD algorithm can restrain mode confusion to some extent, but the effect is not obvious. Therefore, the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) was used to process noisy signals. Firstly, the decomposition results of simulation signals with EMD, EEMD and CEEMDAN were compared, and the permutation entropy values of IMF obtained by EMD, EEMD and CEEMDAN were calculated. Then the decomposition results of EMD, EEMD and CEEMDAN were Hilbert-transformed, and their spectral resolutions were compared. Finally, CEEMDAN was applied to the timefrequency analysis of underwater borehole blasting seismic waves. The results show that CEEMDAN can suppress the mode confusion, and its time-frequency spectrum has high resolution in both time and frequency domain.