基于支持向量机的边坡垂直位移方向率预测及边坡稳定性研究

Support vector machine for vertical displacement direction rate of slope estimate and slope stability research

针对传统边坡位移量预测参数的局限和不足,以及在小样本下的预测参数估计区间较宽会导致工程设计偏于保守的问题,以垂直位移方向率作为边坡稳定性演化分析与评价的一个有效位移动力参数,提出支持向量机模型理论对其在相应周期内的发展趋势进行滚动预测估计,得到周期内垂直位移方向率的预测结果,通过与异常判据值对比得出边坡的稳定程度,并且以京新高速公路胶泥湾路段边坡的实际监测数据为实例进行验证分析。结果表明,预测数据与路基边坡实际情况基本吻合,因此垂直位移方向率可以作为公路边坡失稳判别的有效预测参数,同时支持向量机法评价垂直位移方向率参数对公路路基边坡进行预测预报具有一定的适用性和可靠性。

Since the singular dimension displacement in the slope prediction its limitations,and since the confidence interval of parameters calculated by the traditional method is too wide under the small sample condition which leads to the conservative engineering design,this paper proposes a new appraisal parameter of vertical displacement direction rate as the slope stability,which is an effective displacement dynamic parameter.We combine with the support vector machine model theory to predict its development trend in the corresponding period,and obtain the prediction results of the vertical displacement direction in this period,then contrast with the abnormal criterion value to get the slope stabili⁃ty level.Then taking the actual monitoring data of slopes at Beijing-Xinjiang Highway as an example to carry out the veri⁃fication.The results show that the vertical displacement direction rate can be used as an effective predication parameter of highway slope instability warning,and SVM method to evaluate the vertical displacement direction rate parameter is practically effective in the prediction of highway subgrade slope.