【摘要】 通过对模拟信号的处理，验证了自适应完备经验模态分解(Complete Ensemble Empirical Mode Decomposition with Adaptive Noise, CEEMDAN)方法在资料分析处理过程中的有效性和可靠性。利用该方法对10个GNSS(Global Navigation Satellite System)连续站N、 E、 U三方向时间序列进行分解，发现其不仅存在年周期和半年周期信号，部分台站还存在更短或更长周期信号，且垂向周期性比水平向更明显，对应幅值更大。经周期项改正后，相比于原始N、 E、 U方向时间序列的平均WRMS(Weighted Root Mean Square error)分别降低了19.8%、 19.0%和29.9%,表明该方法对周期项修正的有效性。更多还原

【Abstract】 The validity and reliability of the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise(CEEMDAN) method in the data analysis and processing are verified by processing the simulated signals. Using this method to decompose the N, E and U direction time series of 10 GNSS(Global Navigation Satellite System) continuous stations, it is found that not only annual and semi-annual period signals exist, but also shorter or longer period signals exist in some stations, and the vertical periodicity is more obvious than the horizontal direction, and the corresponding amplitudes are larger. The average WRMS(Weighted Root Mean Square error) of the time series in the N, E, and U directions was reduced by 19.8%, 19.0%, and 29.9%, respectively, compared to the original N, E, and U directions after the periodic term correction, indicating the effectiveness of the method for periodic term correction.更多还原