Preconditioned Conjugate Gradient Methods for the 3D Wilson Nonconforming FEM Discretizations
-
摘要: 非协调元方法是克服三维弹性问题体积闭锁的一种有效方法,它具有自由度少、精度高等优点,但要提高其有限元分析的整体效率还必须为相应的离散化系统设计快速求解算法.考虑了Wilson元离散化系统的快速求解.当Poisson(泊松)比ν→0.5时,该离散系统为一高度病态的正定方程组,预处理共轭梯度(PCG)法是求解这类方程组最为有效的方法之一.另外,在实际应用中,由于结构的特殊性,网格剖分时常常会产生具有大长宽比的各向异性网格,这也将大大影响PCG法的收敛性.该文设计了一种基于“距离矩阵”的代数多重网格(DAMG)法的PCG法,并应用于近不可压缩问题Wilson元离散系统的求解.这种基于“距离矩阵”的代数多重网格法,能更有效地求解各向异性网格问题,再结合有效的磨光算子,相应的PCG法对求解近不可压缩问题具有很好的鲁棒性(robustness)和高效性.
-
关键词:
- 体积闭锁 /
- Wilson非协调元 /
- 病态矩阵 /
- 代数多重网格法 /
- 预条件子
Abstract: The nonconforming finite element method (FEM) is an efficient method to overcome the volume locking trouble in 3D elasticity problems. This method has the advantages of a few degrees of freedom and high accuracy. In order to improve the overall efficiency of the FEM analysis, it is necessary to design some faster solvers for the corresponding system of discretization equations. The faster solvers for the Wilson nonconforming FEM discretizations were considered. When Poisson’s ratio ν was close to 0.5, the resulting system of equations was symmetric positive definite and highly illconditioned, and the preconditioned conjugate gradient (PCG) method was one of the most efficient methods for solving such FEM equations. Moreover, in practical applications, anisotropic meshes are often obtained due to the specificity of the structure considered, which will greatly decrease the convergence rate of the PCG method. A type of PCG method based on the DAMG was presented and then applied to the solution of the Wilson FEM discretizations. This DAMG was an algebraic multi grid (AMG) method based on the distance matrix and can be used to solve the system of equations discretized on anisotropic meshes. The numerical results show that, in combination with the effective smoothing operators, the proposed PCG method has high efficiency and robustness for nearly incompressible problems. -
[1] Morley M. A family of mixed finite elements for linear elasticity[J].Numerische Mathematik,1989,55(6): 633-666. [2] Falk R S. Nonconforming finite element methods for the equations of linear elasticity[J].Mathematics of Computation,1991,57(196): 529-550. [3] QI He, WANG Lie-heng, ZHENG Wei-ying. On locking-free finite element schemes for three dimensional elasticity[J].Journal of Computational Mathematics,2005,23(1): 101-112. [4] Brenner S C. A nonconforming mixed multigrid method for the pure traction problem in planar linear elasticity[J].Mathematics of Computation,1994,63(208): 435-460. [5] Scott L R, Vogelius M. Conforming finite element methods for incompressible and nearly incompressible continua[C]//Large Scale Computations in Fluid Mechanics,Part 2, 1983: 221-244. [6] Hackbusch W.Multi-Grid Methods and Applications [M]. New York: Springer-Verlag, 1985. [7] Trottenberg U, Oosterlee C W, Schuler A.Multigrid[M]. Academic Press, 2001. [8] 刘石, 陈德祥, 冯永新, 徐自力, 郑李坤. 等几何分析的多重网格共轭梯度法[J]. 应用数学和力学,2014,35(6): 630-639.(LIU Shi, CHEN De-xiang, FENG Yong-xin, XU Zi-li, ZHENG Li-kun. A multigrid preconditioned conjugate gradient method for isogeometric analysis[J].Applied Mathematics and Mechanics,2014,35(6): 630-639.(in Chinese)) [9] CHEN Zhang-xin, Oswald P. Multigrid and multilevel methods for nonconforming Q1 element[J].Mathematics of Computation,1998,67(222): 667-693. [10] 周叔子, 文承标. 抛物问题非协调元多重网格法[J]. 计算数学, 1994,16(4): 372-381.(ZHOU Shu-zi, WEN Cheng-biao. Nonconforming element multigrid method for parabolic equations[J].Mathematica Numerica Sinica,1994,16(4): 372-381.(in Chinese)) [11] 许学军. 旋转Q1非协调元的V循环多重网格法[J] . 计算数学,1999,21(2): 251-256.(XU Xue-jun. V-cycle multigrid methods for Q1 nonconforming finite elements[J].Mathematica Numerica Sinica,1999,21(2): 251-256.(in Chinese)) [12] SHI Zhong-ci, XU Xue-jun. V-cycle multigrid methods for Wilson nonconforming element[J].Science in China(Series A): Mathematics,2000,43(7): 673-684. [13] Chartier T P, Falgout R D, Henson V E, Jones J E, Maneuffel T, McCormick S F, Ruge J W, Vassilevski P S. Spectral AMGe(ρAMGe)[J].SIAM Journal on Scientific Computing,2003,20(1): 1-20. [14] Brezina M, Cleary A J, Falgout R D, Henson V E, Jones J E, Manteuffel T A, McCormick S F, Ruge J W. Algebraic multigrid based on element interpolation (AMGe)[J].SIAM Journal on Scientific Computing,2000,22(5): 1570-1592. [15] Kolev T V, Vassilevski P S. AMG by element agglomeration and constrained energy minimization interpolation[J].Numerical Linear Algebra With Applications,2006,13(9): 771-788. [16] XIAO Ying-xiong, SHU Shi, ZHANG Ping, TAN Min. An algebraic multigrid method for isotropic linear elasticity on anisotropic meshes[J].International Journal for Numerical Methods in Biomedical Engineering,2010,26(5): 534-553. [17] El maliki A, Guénette R, Fortin M. An efficient hierarchical preconditioner for quadratic discretizations of finite element problems[J].Numerical Linear Algebra With Applications,2011,18(5): 789-803. [18] XIAO Ying-xiong, SHU Shi, ZHAO Tu-yan. A geometric-based algebraic multigrid for higher-order finite element equations in two dimensional linear elasticity[J].Numerical Linear Algebra With Applications,2009,16(7): 535-559. [19] Saad Y.Iterative Methods for Sparse Linear Systems[M]. 2nd ed. Philadelphia, PA: Society for Industrial and Applied Mathematics, 2003. [20] Notay Y. An aggregation-based algebraic multigrid method[J].Institute of Computational Mathematics,2010,37: 123-146.
点击查看大图
计量
- 文章访问数: 1239
- HTML全文浏览量: 162
- PDF下载量: 549
- 被引次数: 0