Numerical analysis of a least-squares finite element method for the time-dependent advection-diffusion equation (Q534247): Difference between revisions

From MaRDI portal
RedirectionBot (talk | contribs)
Removed claim: author (P16): Item:Q459072
RedirectionBot (talk | contribs)
Changed an Item
Property / author
 
Property / author: Vitoriano Ruas / rank
 
Normal rank

Revision as of 03:32, 15 February 2024

scientific article
Language Label Description Also known as
English
Numerical analysis of a least-squares finite element method for the time-dependent advection-diffusion equation
scientific article

    Statements

    Numerical analysis of a least-squares finite element method for the time-dependent advection-diffusion equation (English)
    0 references
    0 references
    0 references
    17 May 2011
    0 references
    The authors consider the equation named in the title in a bounded \(N\)-dimensional domain along with initial and mixed Dirichlet and Neumann boundary values. For its numerical solution they apply the Crank-Nicolson scheme in time and mixed finite elements (approximating the solution and its flux by piecewise polynomials of possibly different degrees) in space, combined with a least squares approach containing a free parameter. They prove unicity of the discrete solution, and, under conditions on the least squares parameter and on the time step (the latter depending on the Péclet number) the stability. Then convergence is shown for the case that \(h\) is sufficiently small and that the time step is connected to a power of \(h\). Numerical experiments (using equal-degree approximations for solution and flux) show the agreement of the theoretical results for the convergence order of the gradient of the solution, but the order of convergence of the solution and the flux turned out to be somewhat greater in the unit-square example.
    0 references
    advection-diffusion-reaction equation
    0 references
    Crank-Nicolson scheme
    0 references
    mixed Lagrangian finite elements
    0 references
    least squares approach
    0 references
    convergence
    0 references
    stability
    0 references
    numerical experiments
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references