基于支持向量机的土石坝边坡稳定概率分析

Probability analysis of slope failure of embankment slope based on support vector machine

为了合理考虑土石坝中不确定性因素的影响,本文基于支持向量机方法进行了土石坝边坡稳定概率分析,详细描述了整体计算流程,并通过一个土石坝边坡算例验证了该方法的有效性。结果表明:该方法可以结合参数归一化方法及“0”和“1”输出项来提升土石坝边坡分类精度,仅需进行30次土石坝边坡渗流稳定分析(耗时5min),便可有效替代1万次MCS方法(耗时20h)进行土石坝边坡概率分析。该方法计算效率高、耗时少,可为解决考虑不确定性的土石坝边坡稳定概率分析问题提供一条有效的途径。

In order to reasonably consider the influence of uncertainties of embankment slope,based on the support vector machine method,this paper conducts the probabilistic analysis of slope failure of embankment.The detailed calculation flow is illustrated,and the effectiveness of the proposed method is verified by a case of embankment slope.The results show that the proposed method combines the parameter normalization method and the output terms of"0"and"1"to improve the classification accuracy of embankment slope.The proposed method only needs 30 times of seepage stability analysis of embankment slope spending 5 min,which can effectively replace 10,000 times of MCS method spending 20 h to make the probability analysis of embankment slope.The proposed method has high calculation efficiency and less time consumption,which provides an effective tool for solving the probability analysis of embankment slope considering soil uncertainties.