边坡可靠度分析中岩土合理样本数目探讨

Study on the geotechnical number of samples in slope reliability analysis

土工参数由有限样本数目勘察取样试验结果的估计推断而得到,统计不确定性将对可靠度分析结果带来误差,因此合理样本数目应体现分析结果精度与勘察费用间矛盾的协调。基于从土工参数总体分布中随机抽样的方法模拟某一样本容量下的勘察取样试验,以120次重复抽样构建样本统计值空间,分析样本统计值及对应边坡可靠度指标的离散特征。根据样本容量对可靠度指标计算结果空间散布范围的约束效应,探讨土工强度参数变异水平呈小、中和大3个等级时相对应的合理样本数目。计算分析表明:样本数目增加可减小可靠指标分布的离散程度,土工参数变异性增大则反之;为保证可靠度分析结果精度,在土工参数变异性增大条件下须相应增加样本数目;可靠度指标变异系数取0.30时,对应于参数呈小、中和大变异水平的样本数目分别为6,11和18。研究成果对完善工程勘察中的土工试验取样数目有一定参考价值。

Soil parameters were estimated by the survey test of limited sample number and its statistical uncertainty would lead to the error of reliability analysis.Thus a number of samples should be considered in terms of survey cost and precision.Based on the method of random sampling from the total soil parameters to simulate the survey under certain sample number,the sample statistical value and the discrete features of slope reliability index correspondingly are analyzed with 120 times sampling.Based on the influence of sample number on discrete features of slope reliability index,the number of sample corresponding to three levels of soil parameters is discussed.The analysis result shows that the increased number of samples would lead to the small dispersion degree of reliability index,and the increasing variability of parameters would tead to the opposite result.In order to ensure the accuracy of the results,the number of sample should increase with the increase of reliability index of soil parameters.According to small,medium and large variation levels of soil parameters,the required number of samples are found to be 6,11 and 18 seperately when the coefficient of variation is 0.30.Research provides reference on the soil test of engineering survey.