基于LHS抽样RSM数据表的边坡可靠度研究

为有效求解边坡可靠度,采用拉丁超立方(LHS)抽样提出了基于响应面法(RSM)数据表的边坡可靠度计算模型。考虑影响边坡稳定性的主要强度参数c,φ的不确定性,采用LHS抽样构建RSM随机样本点,借助岩土工程极限平衡的Slide程序获取样本响应值;通过将数据表法与RSM结合求解边坡可靠指标及失效概率。以Monte Carlo法计算结果作为可靠性分析的基准解,并与基本2层土坡模型和托巴110 kV施工变电站边坡计算结果进行对比分析,结果表明此方法不仅精度满足要求,且计算量较小、效率高,对实际工程边坡可靠度分析具有一定指导意义。

Slope Reliability Research Based on Response Surface Method withSpreadsheeet of Latin Hypercube Sampling

A slope reliability calculation model combining spreadsheet method and Response Surface Method (RSM) based on Latin Hypercube Sampling (LHS) is presented to solve slope reliability issue effectively. In consideration of the uncertainties of strength parameters cohesion c and friction angle φ which have significant impact on slope stability, LHS sampling is used to establish random sample points of RSM, and the response values of sample points are obtained based on slide program. By combining the spreadsheet method with the RSM, the reliability index and failure probabilities are solved. With Monte Carlo results as a benchmark reference, a comparison of calculation results for basic two-layer slope model and Tuoba 110 kV substation slope show that the proposed method is of high accuracy, less computational efforts, and higher efficiency, and is of guiding importance for the reliability analysis of practical engineering slope.