摘要
提出了基于稀疏网格配点法的边坡可靠度分析方法;利用有限元强度折减法计算了边坡安全系数,再通过稀疏网格配点法构造代理模型,将隐式功能函数转为显式功能函数,从而降低了有限元计算次数。在此基础上结合蒙特卡罗方法计算边坡失效概率,给出了计算流程,并编写了相应的MATLAB程序。最后以一显式功能函数问题为例验证了稀疏网格配点法的有效性,并以两个边坡问题为例研究了所提方法在边坡可靠度中的应用。结果表明,虽然1阶Smolyak配点法相比2阶Smolyak配点法计算量小,但后者更适用于边坡可靠度分析。所提方法能够对考虑土体参数空间变异性的边坡可靠度进行分析,计算结果与拉丁超立方抽样方法保持一致,计算精度较高,为进行复杂的边坡可靠度分析提供了一条有效的途径。
A new methodology is proposed for reliability analysis of slope stability based on the sparse grid stochastic collocation method.The factor of safety is calculated by using the strength reduction method.The proposed method can well fit the actual performance function for slope reliability analysis using a surrogate model, which has an excellent modeling capability.The probability of slope failure is then estimated using Monte-Carlo simulation based on this explicit performance function.A computer program is developed.An explicit limit-state function-example is provided to validate the accuracy and efficiency of this method.Moreover,taking the two soil slope problems as examples,the application of the proposed method in slope reliability is studied.Results indicate that although the first-order Smolyak collocation method is less computationally intensive than the second-order Smolyak collocation method,the latter is more suitable for slope reliability analysis.The proposed method can analyze the slope reliability by considering the spatial variability of soil parameters,and the calculated results are consistent with the Latin hypercube sampling method,which has high calculation accuracy and provides an effective way to perform complex slope reliability analysis.
作者
潘敏
凌晨
范晶晶
牛景太
Pan Min;Ling Chen;Fan Jingjing;Niu Jingtai(School of Hydraulic and Ecological Engineering,Nanchang Institute of Technology,330099,Nanchang,China;Zhejiang Design Institute of Water Conservancy &Hydro-Electric Power,310002,Hangzhou,China)
出处
《应用力学学报》
CAS
CSCD
北大核心
2018年第6期1267-1272,1419,1420,共8页
Chinese Journal of Applied Mechanics
基金
国家自然科学基金地区科学基金项目(51769017)
浙江省水利厅科技计划项目(RC1546)
关键词
边坡
可靠度
稀疏网格配点
土体参数
不确定性
slope
reliability
sparse grid stochastic collocation
soil parameters
uncertainty
