Propagation and spurious reflection in finite-element approximations of hyperbolic equations (Q1083181)

From MaRDI portal
Revision as of 17:16, 19 February 2024 by RedirectionBot (talk | contribs) (‎Removed claims)
scientific article
Language Label Description Also known as
English
Propagation and spurious reflection in finite-element approximations of hyperbolic equations
scientific article

    Statements

    Propagation and spurious reflection in finite-element approximations of hyperbolic equations (English)
    0 references
    1985
    0 references
    The author considers as a model problem the simple hyperbolic equation (1) \(\partial u/\partial t+c \partial u/\partial x=0\), \((c>0)\) approximated by a linear finite element Galerkin method on a regular net. The corresponding fully discrete approximation of (1) is the implicit system \(M(u^ j_{n-1}+4u^ j_ n-u^ j_{n+1})/6=-C(u^ j_{n+1}-u^ j_{n-1})/2h\) where \(M(z)=2(z-1)/\Delta t(z+1)\) is the Crank-Nicolson time-shift operator. For this approximation the author shows that the energy conserved by this scheme is not the square of the \(1_ 2\) norm as it is with finite differences. Moreover the group velocities of rightgoing and leftgoing solutions are not equal. Illustrations of spurious reflection at downwind boundaries, upwind boundaries and at interfaces in mesh refinement are given together with their analysis.
    0 references
    spurious solution
    0 references
    energy conservation
    0 references
    finite element Galerkin
    0 references
    Crank- Nicolson
    0 references
    group velocities
    0 references
    spurious reflection
    0 references

    Identifiers