边界(拓扑)
插值(计算机图形学)
网格
移动最小二乘法
数学
面子(社会学概念)
引力奇点
复杂几何
领域(数学分析)
歧管(流体力学)
正则化无网格法
数值分析
网格生成
应用数学
几何学
数学优化
算法
计算机科学
奇异边界法
数学分析
人工智能
边界元法
图像(数学)
有限元法
结构工程
工程类
机械工程
社会学
社会科学
作者
Shan Lin,Xitailang Cao,Hong Zheng,Yanyan Li,Wei Li
标识
DOI:10.1016/j.compgeo.2022.105211
摘要
Accurate seepage calculation is a prerequisite to ensuring the safety of water projects. But seepage problems face low accuracy and mesh dependence when conventional methods simulate the complex boundaries. Here, complex boundaries are reflected in the complexity of solving problems with multiple types of singularities and the complexity of the geometry of the problem domain. This paper uses the numerical manifold method based on moving least squares (MLS-NMM) interpolation to solve seepage problems with complex boundaries. The method adopts two cover systems and discrete node-based interpolation, so it owns significant advantages in dealing with complex boundaries. In order to improve the accuracy for simulating the corner singularities and jump intermittent points of essential boundaries, an improved MLS-NMM is established. Simultaneously, a background integral grid generation method based on the physical cover is introduced to improve computational accuracy and simplify the pre-processing of complex geometric problem domains, which also acquires a simple and regular background integral grid. A seepage problem with the free surface can be simulated with high accuracy by the improved MLS-NMM. The calculation results of complex boundary seepage cases with multiple types of singularities are compared with the exact or reference solutions, confirming the proposed method's validity and good applicability. As a result, the improved MLS-NMM can be used as a reference tool for analyzing complex seepage in engineering.
科研通智能强力驱动
Strongly Powered by AbleSci AI