论文部分内容阅读
为准确预测滑坡深部位移,以福利院滑坡为例,选择小波去噪处理变形监测数据,并探讨了小波函数、小波分解层数、阈值选择标准和阈值选取方法对小波去噪效果的影响,发现采用sym7小波函数、5层小波分解、启发式阈值和软阈值选取阈值时去噪效果最好,进而对去噪后的数据进行回归分析,对比分析了不同回归函数的效果。结果表明,傅里叶回归函数的回归效果最优,因而利用傅里叶回归函数预测了滑坡深部位移的变形趋势,为判断滑坡的稳定性及治理提供了依据。
In order to accurately forecast the deep displacement of landslide, taking Wuliyuan landslide as an example, the wavelet de-noising processing deformation monitoring data is selected, and the influence of wavelet function, wavelet decomposition level, threshold selection criterion and threshold selection method on wavelet denoising effect is investigated. Using sym7 wavelet function, 5-layer wavelet decomposition, heuristic threshold and soft threshold to select the threshold value, the denoising effect is the best, and then the regression analysis of the data after denoising, comparative analysis of the effect of different regression functions. The results show that the regression function of Fourier regression function is the best. Therefore, the deformation trend of the deep displacement of landslide is predicted by using the Fourier regression function, which provides a basis for judging the stability and treatment of the landslide.