中文摘要：

目前在滑坡位移预测研究中,先将滑坡位移数据分解为趋势项及周期项后再分别进行预测已成为普遍做法。平滑先验分析法（Smoothness Priors Approach,SPA）是一种计算过程简单、计算量极小,且能快速分离原始数据趋势项和周期项的数据处理方法。在介绍SPA基本原理基础上,以三峡库区白家包滑坡典型监测点位移数据为例,对通过调节SPA正则化参数而获得的不同趋势项及周期项进行特征分析;进而结合对滑坡变形演化机制过程的先验分析,根据位移分解特征确定合理的参数取值;最后针对不同参数SPA位移分解数据,采用支持向量机进行位移预测对比分析。结果表明,SPA是一种适用于滑坡位移预测的位移分解方法,通过调节正则化参数并结合滑坡变形机制先验分析,能够获得较为合理的位移分解结果,进而提高滑坡位移预测精度。

英文摘要：

At present,in landslide displacement prediction,it has become a widespread practice that the original displacement is first decomposed into trend term and periodic term. Smoothness Priors Approach（ SPA） is such a data decomposition method and has some advantages including simple calculation process,minimal computational effort,etc. Based on the introduction of the basic principle of the SPA,the typical displacement monitoring data of the Baijiabao landslide in the Three Gorges reservoir area is used to analyze the characteristics of trend and periodic term displacement according to different regularization parameter of SPA. The results show that the trend term is smoother,the amplitude is larger and the period is longer with the increase of parameters. Through the prior analysis of the landslide deformation mechanism and processes,the most reasonable parameters can be determined. Finally,landslide displacement prediction is carried out for different displacement decomposition data,which is obtained from different parameters of SPA and with the moving average method. The results show that there is very good prediction accuracy based on the SPA displacement decomposition. Comprehensive analysis indicates that the SPA is a suitable method for the displacement decomposition in the landslide displacement prediction,and the prediction accuracy can further be improved.