Duncan法在黄土高边坡可靠度分析中的应用研究

Application of Duncan Method in Reliability Analysis of Loess High Slope

以4种典型的黄土高边坡为例,通过Geo-slope软件建立地质模型,采用Monte-Carlo法和Duncan法对土体参数变异情况下边坡进行可靠度分析,并以Monte-Carlo法模拟10 000次的计算结果为准确值,对比分析了Duncan法的精确度,引入相对误差修正Duncan法的计算结果。结果表明,在不同参数变异系数组合下,Duncan法估计的黄土边坡失效概率高于Monte-Carlo法计算结果,绝对误差在5%以内,计算结果相近,参数的变异性对2种方法的计算误差影响小;2种方法的相对误差呈对数线性关系。由此,提出一种相对于Monte-Carlo法的失效概率相对误差计算公式,对Duncan法计算的失效概率进行修正。Duncan法修正后计算的边坡失效概率理论简单,计算量小,适应于实际黄土边坡可靠度分析。

In this paper,four typical loess high slopes are taken as examples,and Geo-slope software is used to establish a geological model.Monte-Carlo method and Duncan method are adopted to analyze the reliability of the slope under the variation of soil parameters.The calculation results of the simulation of 10,000 times by Monte-Carlo method are taken as the accurate value,and the accuracy of Duncan method is compared and analyzed.The results show that the failure probability of loess slope estimated by Duncan method is higher than that calculated by Monte-Carlo method under the combination of variation coefficients of different parameters.The absolute error is within 5%,and the influence of parameter variability on the absolute error is slight.The relative errors of the two methods are logarithmically linear.Therefore,a formula for calculating the relative error of failure probability compared with Monte-Carlo method is proposed.Then the failure probability calculated by Duncan method is modified.The modified Duncan method is simple in theory and small in calculation,which is more suitable for the reliability analysis of actual loess slope.