Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
520229 | Journal of Computational Physics | 2012 | 21 Pages |
Abstract
The paper develops a posteriori error estimates of integral output functionals for summation-by-parts finite-difference methods. The error estimates are based on the adjoint-weighted residual method and take advantage of a variational interpretation of summation-by-parts discretizations. The estimates are computed on a fixed grid and do not require an embedded grid or explicit interpolation operators. For smooth boundary-value problems containing first and second derivatives the error estimates converge to the exact error as the mesh is refined. The theory is verified using linear boundary-value problems and the Euler equations.
Related Topics
Physical Sciences and Engineering
Computer Science
Computer Science Applications
Authors
J.E. Hicken,