考虑尾矿材料参数空间变异性的尾矿坝稳定可靠度分析

为客观地评价尾矿坝稳定性，在随机场理论和极限平衡分析框架下，采用考虑参数空间变异性尾矿坝可靠度分析的非侵入式随机有限元法，通过Karhunen Loève级数展开方法离散尾矿材料参数（渗透系数、摩擦角等）随机场，利用Hermite随机多项式展开拟合尾矿坝安全系数与输入参数之间的隐式函数关系；再采用拉丁超立方抽样技术产生输入参数样本点求解多项式展开系数；最后，通过应用到一实际尾矿坝工程说明了该方法的有效性。结果表明，该方法可以真实地模拟尾矿材料的空间变异性对尾矿坝稳定性的影响，保证尾矿坝可靠度分析与确定性稳定分析互不耦合，与10 000次直接拉丁超立方抽样方法相比，该方法具有较高的计算精度和效率。

Reliability analysis of tailings dam stability considering spatial variability of tailings materials parameters

In order to evaluate the stability of tailings dam objectively,a non intrusive stochastic finite element method considering the spatial variability of tailings materials parameters was developed for the reliability analysis of tailings dam under the framework of random field theory and limit equilibrium analysis.The Karhunen Loève series expansion method was adopted for the discretization of random fields of tailings materials parameters (e.g.,permeability coefficient and friction angle).The Hermite polynomial chaos expansion (HPCE) was utilized to fit the implicit function relationship between the safety coefficient of tailings dam and the input parameters,and the Latin hypercube sampling technology was employed to generate the sample points of input parameters for determining the coefficients of HPCE.Finally,the effectiveness of the proposed method was demonstrated by applying in an actual tailings dam engineering.The results showed that the proposed method could truly simulate the effect of the spatial variability of tailings materials parameters on the stability of tailings dam,and ensure the mutual non coupling of the reliability analysis and the deterministic stability analysis of the tailings dam.In comparison with the Latin hypercube sampling method with a size of 10,000 samples,the proposed method achieved higher calculation accuracy and efficiency.