求解无限域动力刚度矩阵的双渐近算法

A DOUBLY ASYMPTOTIC ALGORITHM FOR SOLVING THE DYNAMIC STIFFNESS MATRIX OVER UNBOUNDED DOMAINS

  • 摘要: 采用连分式算法可以有效地求解无限域动力刚度表示的比例边界有限元方程, 它具有收敛范围广、收敛速度快等优点. 该文在高频渐近连分式算法的基础上考虑了低频渐近, 发展了一种针对矢量波动方程的双渐近算法. 随着展开阶数的增加, 双渐近算法可以在全频域范围内快速逼近准确解. 引入了系数矩阵?X(i)来增强连分式算法的数值稳定性. 通过在高频极限、低频极限时满足动力刚度表示的比例边界有限元方程, 建立了递推关系以求得动力刚度矩阵. 通过二维半无限楔形体、三维均质弹性半空间数值算例表明, 双渐近算法比单渐近算法更稳定、优越.

     

    Abstract: The continued-fraction algorithm can efficiently solve a scaled boundary finite element equation expressing the dynamic stiffness over an unbounded domain, and has a large convergence range and a high convergence rate. In this paper, a doubly asymptotic continued-fraction algorithm based on the singly asymptotic approach is developed to solve vector wave equations; the algorithm rapidly converges to the exact solution over the whole frequency range with increasing order of expansion. The factor matrices X(i) are introduced to the continued fraction solution to improve the numerical stability of the solution. The coefficients of the solution are determined recursively by satisfying the scaled boundary finite element equation for dynamic stiffness at both the high and low frequency limits. The solutions for a 2-D semi-infinite wedge and a 3-D homogeneous half space are presented as examples. The results of these two examples demonstrate that the doubly asymptotic algorithm is more stable than and superior to the singly asymptotic approach.

     

/

返回文章
返回