基于小波分解的高边坡变形预报

Based on wavelet decomposition of the high slope deformation prediction

指出小波多分辨分析具有较强的时频分析特性,对含有趋势性、周期性和随机性的非线性变形时间序列进行分解,用不同的模型对各分解项进行预测后叠加,比单纯用某一种模型对变形的预测,精度有较大的提高,对各分解项用时间序列分析模型预测周期性,用多项式拟合趋势性,结果显示,预测效果较好。

Points out multi-resolution and analysis of wavelet analysis has strong characteristics of time-frequency analysis, it can resolue the deformation time series which have trend cyclical and non-linear, use the different modeles to predict the decomposition and overlay, the forecast accuracy has been enhanced to simply use a one model. In this paper, to the decomposition use the time series analysis models to predict cyclical, and use the polynomial fitting with the trend. The results show that the forecast is better than the traditional method.