奇异谱分析在倾斜应变数据处理中的应用研究

通过模拟数据实验验证奇异谱分析（SSA）对于周期信号和趋势项有分离和提取功能，能有效去除标准差低于信号振幅的噪声，检测识别信号中存在的奇异点。将SSA应用于实际倾斜应变固体潮数据的处理中，选取丹东台、徐州台和泾县台的倾斜应变数据进行SSA处理。结果表明，SSA能有效提取数据中的长趋势成分和固体潮各潮波，有效分离数据中的干扰成分，识别出周期为7~15 d的降雨干扰, 对干扰排除和异常提取具有一定的参考意义。最后经过大量数据验证得出，SSA适用于倾斜应变数据的嵌入维度范围的经验选择。

Research on Application of Singular Spectrum Analysis in Data Processing of Earth Tilt and Strain

The results of simulation experiments show that SSA can separate and extract the periodic signal and trends, effectively reduce noise of which the standard deviation is lower than the signal amplitude, and detect singular points in signals. Through applying SSA to data processing of earth tilt and strain in Dandong, Xuzhou and Jingxian stations, the results indicate that SSA can extract components of long trends of the data, separate tidal waves effectively, and identify the rain interference components of which the periodic signal is about 7 to 15 days. These results have some reference implications for disturbing rejection and extracting exception. Finally, after significant data validation, we summarize choice of the experience embedding dimension when SSA is used for tilt and strain data.