高级检索

基于均值坐标的多边形单元r细化方法

A Method for r Refinement of Polygonal Finite Element Based on Mean Value Coordinates

  • 摘要: 局部和各向异性特征广泛地存在于各种物理现象中. 为提高该现象建模方程的求解精度, 提出了一种基于均值坐标的多边形单元r细化方法. 利用二阶均值坐标形函数与均值坐标形函数下解的误差构造目标函数; 通过优化多边形单元种子点的位置, 迭代更新网格及其方程的解曲面; 促使曲面特征变换剧烈的地方具有较多的自由度. 求解二维Poisson方程的数值实例结果表明, 在相同种子点数量下, 各项误差可以减小50%以上, 所提方法能有效地提高解的精度.

     

    Abstract: Local and anisotropic features are widely present in various physical phenomena. In order to enhance the accuracy of modeling equations for this phenomenon, a r-refinement method based on mean value coordinates for polygonal elements is proposed. The objective function is constructed using the error between the solution under the mean value coordinates based on quadratic serendipity element shape functions and the mean value coordinates shape functions. By optimizing the locations of sites in polygonal elements, the mesh is iteratively updated along with the solution surface of the equation, allowing regions with drastic surface characteristic changes to have more degrees of freedom. Numerical results of solving the 2D Poisson equation demonstrate that with the same number of the sites, the error can be reduced by over 50%, indicating that the proposed method effectively improves the accuracy of the solution.

     

/

返回文章
返回