Loading...

Table of Content

    15 September 2012, Volume 30 Issue 5
    THE BEST L2 NORM ERROR ESTIMATE OF LOWER ORDERFINITE ELEMENT METHODS FOR THE FOURTH ORDERPROBLEM
    Jun Hu, Zhong-Ci Shi
    2012, 30(5):  449-460.  DOI: 10.4208/jcm.1203-m3855
    Asbtract ( 400 )   PDF  
    References | Related Articles | Metrics
    In the paper, we analyze the L2 norm error estimate of lower order finite elementmethods for the fourth order problem. We prove that the best error estimate in the L2norm of the finite element solution is of second order, which can not be improved generally.The main ingredients are the saturation condition established for these elements and anidentity for the error in the energy norm of the finite element solution. The result holdsfor most of the popular lower order finite element methods in the literature including: thePowell-Sabin C1-P2 macro element, the nonconforming Morley element, the C1-Q2 macroelement, the nonconforming rectangle Morley element, and the nonconforming incompletebiquadratic element. In addition, the result actually applies to the nonconforming Adinielement, the nonconforming Fraeijs de Veubeke elements, and the nonconforming Wang-Xu element and the Wang-Shi-Xu element provided that the saturation condition holdsfor them. This result solves one long standing problem in the literature: can the L2 normerror estimate of lower order finite element methods of the fourth order problem be twoorder higher than the error estimate in the energy norm?
    PRECONDITIONING THE INCOMPRESSIBLE NAVIER-STOKESEQUATIONS WITH VARIABLE VISCOSITY
    Xin He, Maya Neytcheva
    2012, 30(5):  461-482.  DOI: 10.4208/jcm.1201-m3848
    Asbtract ( 509 )   PDF  
    References | Related Articles | Metrics
    This paper deals with preconditioners for the iterative solution of the discrete Oseenproblem with variable viscosity. The motivation of this work originates from numericalsimulations of multiphase flow, governed by the coupled Cahn-Hilliard and incompressibleNavier-Stokes equations. The impact of variable viscosity on some known preconditioningtechnique is analyzed. Theoretical considerations and numerical experiments showthat some broadly used preconditioning techniques for the Oseen problem with constantviscosity are also effcient when the viscosity is varying.
    ERROR REDUCTION, CONVERGENCE AND OPTIMALITYFOR ADAPTIVE MIXED FINITE ELEMENT METHODS FORDIFFUSION EQUATIONS
    Shaohong Du, Xiaoping Xie
    2012, 30(5):  483-503.  DOI: 10.4208/jcm.1112-m3480
    Asbtract ( 426 )   PDF  
    References | Related Articles | Metrics
    Error reduction, convergence and optimality are analyzed for adaptive mixed finiteelement methods (AMFEM) for diffusion equations without marking the oscillation ofdata. Firstly, the quasi-error, i.e. the sum of the stress variable error and the scaled errorestimator, is shown to reduce with a fixed factor between two successive adaptive loops,up to an oscillation. Secondly, the convergence of AMFEM is obtained with respect to thequasi-error plus the divergence of the flux error. Finally, the quasi-optimal convergencerate is established for the total error, i.e. the stress variable error plus the data oscillation.
    A NUMERICAL METHOD FOR SOLVING THE ELLIPTICINTERFACE PROBLEMS WITH MULTI-DOMAINS ANDTRIPLE JUNCTION POINTS
    Songming Hou, Liqun Wang, Wei Wang
    2012, 30(5):  504-516.  DOI: 10.4208/jcm.1203-m3725
    Asbtract ( 462 )   PDF  
    References | Related Articles | Metrics
    Elliptic interface problems with multi-domains and triple junction points have wideapplications in engineering and science. However, the corner singularity makes it a challengingproblem for most existing methods. An accurate and efficient method is desired. Inthis paper, an efficient non-traditional finite element method with non-body-fitting grids isproposed to solve the elliptic interface problems with multi-domains and triple junctions.The resulting linear system of equations is positive definite if the matrix coefficients forthe elliptic equations in the domains are positive definite. Numerical experiments showthat this method is about second order accurate in the L norm for piecewise smoothsolutions. Corner singularity can be handled in a way such that the accuracy does notdegenerate. The triple junction is carefully resolved and it does not need to be placedon the grid, giving our method the potential to treat moving interface problems withoutregenerating mesh.
    STABILITY AND RESONANCES OF MULTISTEP COSINEMETHODS
    B. Cano, M.J. Moreta
    2012, 30(5):  517-532.  DOI: 10.4208/jcm.1203-m3487
    Asbtract ( 383 )   PDF  
    References | Related Articles | Metrics
    In a previous paper, some particular multistep cosine methods were constructed whichproved to be very efficient because of being able to integrate in a stable and explicit waylinearly stiff problems of second-order in time. In the present paper, the conditions whichguarantee stability for general methods of this type are given, as well as a thorough studyof resonances and filtering for symmetric ones (which, in another paper, have been provedto behave very advantageously with respect to conservation of invariants in Hamiltonianwave equations). What is given here is a systematic way to analyse and treat any of themethods of this type in the mentioned aspects.
    BANDED TOEPLITZ PRECONDITIONERS FOR TOEPLITZMATRICES FROM SINC METHODS
    Zhi-Ru Ren
    2012, 30(5):  533-543.  DOI: 10.4208/jcm.1203-m3761
    Asbtract ( 391 )   PDF  
    References | Related Articles | Metrics
    We give general expressions, analyze algebraic properties and derive eigenvalue boundsfor a sequence of Toeplitz matrices associated with the sinc discretizations of various ordersof differential operators. We demonstrate that these Toeplitz matrices can be satisfactorilypreconditioned by certain banded Toeplitz matrices through showing that the spectra ofthe preconditioned matrices are uniformly bounded. In particular, we also derive eigenvaluebounds for the banded Toeplitz preconditioners. These results are elementary inconstructing high-quality structured preconditioners for the systems of linear equationsarising from the sinc discretizations of ordinary and partial differential equations, and areuseful in analyzing algebraic properties and deriving eigenvalue bounds for the correspondingpreconditioned matrices. Numerical examples are given to show effectiveness of thebanded Toeplitz preconditioners.
    A P-VERSION TWO LEVEL SPLINE METHOD FORSEMI-LINEAR ELLIPTIC EQUATIONS
    Xinping Shao, Danfu Han, Xianliang Hu
    2012, 30(5):  544-554.  DOI: 10.4208/jcm.1203-m3813
    Asbtract ( 462 )   PDF  
    References | Related Articles | Metrics
    A novel two level spline method is proposed for semi-linear elliptic equations, wherethe two level iteration is implemented between a pair of hierarchical spline spaces withdifferent orders. The new two level method is implementation in a manner of p-adaptivity.A coarse solution is obtained from solving the model problem in the low order spline space,and the solution with higher accuracy are generated subsequently, via one step Newtonor monidifed Newton iteration in the high order spline space. We also derive the optimalerror estimations for the proposed two level schemes. At last, the illustrated numericalresults confirm our error estimations and further research topics are commented.
    THE ULTRACONVERGENCE OF EIGENVALUES FORBI-QUADRATIC FINITE ELEMENTS
    Lingxiong Meng, Zhimin Zhang
    2012, 30(5):  555-564.  DOI: 10.4208/jcm.1203-m3977
    Asbtract ( 416 )   PDF  
    References | Related Articles | Metrics
    The classical eigenvalue problem of the second-order elliptic operator is approximatedwith bi-quadratic nite element in this paper. We construct a new superconvergent functionrecovery operator, from which the O(h8|lnh|2) ultraconvergence of eigenvalue approxima-tion is obtained. Numerical experiments verify the theoretical results.