中文摘要：

工程实践中岩土取样数目通常较少，定量分析样本数量对边坡可靠度分析结果的影响大小，以保证分析结果的可靠性对工程实践有一定参考价值。考虑边坡饱和土体的不固结不排水抗剪强度，从2个角度出发探讨不同岩土样本数目对边坡可靠度分析结果的影响：（1）假设参数的变异性已知，考虑统计不确定性，研究其对边坡可靠度分析结果的影响；（2）提出一种模拟实际边坡工程取样试验的直观分析方法，根据多组参数抽样结果对边坡进行可靠度分析，得到边坡破坏概率的频数分布，进而由实际需要确定所需样本数量。分析结果表明，当参数变异性较大时，所需样本数较少，当参数变异性较小时，所需样本数较多，这与传统的概率方法确定样本数目结果不同。模拟边坡取样试验的分析结果表明：从边坡实际设计上看，对于变异系数为0．2～0．5的不排水强度参数所需样本数目一般在10～15范围内。

英文摘要：

The number of geotechnical samples is usually limited in engineering practice. Evaluating the influence of sample number on slope reliability analysis so as to ensure the reliability of analysis results is of great importance. Considering unconsolidated-undrained shear strength of slope, the influence of the number ofgeotechnical samples on slope reliability is analyzed by two ways： （1） Assuming that the variability of parameters is known and statistical uncertainty is considered in slope reliability analysis. （2） A simple method for simulating geotechnical sampling and testing of actual slope is presented. Slope reliability is computed based on different numbers sampling results, and the frequency distribution of failure probability of slope is obtained, so as to determine the required number of geotechnical samples. The analysis results by the two methods show that when the parameter variability is large, a small number of samples is required; and when the parameter variability is small, large number of samples is required, which differs from the results by classical probabilistic methods. Results from the second method also show that： from the view of practical slope design, the required number of samples for the undrained strength with common coefficient of variation ranging from 0.2 to 0.5, is generally between 10 and 15.