The numerical solution of the Helmholtz equation for wave propagation problems in underwater acoustics (Q1079077)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | The numerical solution of the Helmholtz equation for wave propagation problems in underwater acoustics |
scientific article |
Statements
The numerical solution of the Helmholtz equation for wave propagation problems in underwater acoustics (English)
0 references
1985
0 references
This paper discusses a numerical method for approximately solving the Helmholtz equation with a variable index of refraction in the intermediate frequency range: \(\Delta u+k^ 2n^ 2(x,y)u=0\), where u is the acoustic pressure, n(x,y) is the index of refraction, k is the wave number. The region of interest is a duct or waveguide containing inhomogeneities and interfaces. The Helmholtz equation is discretized using the finite element method, the resulting matrices is indefinite and nonselfadjoint due to the radiation boundary condition. The linear system of equations is solved iteratively by using the preconditioned conjugate gradient method applied to the normal equations. The effectiveness of the iterative method depends on the choice of the preconditioner. The preconditioners are based on a splitting of the discrete Laplacian. In this paper, the preconditioner is constructed based on a multigrid method, it has the additional advantage of being vectorizable, since a relaxation scheme based on a red black ordering is used. The use of multigrid as a preconditioner with a restricted number of levels give a rate of increase of O(k), instead of \(O(k^ 2)\) or \(O(k^{3/2})\) as compared to a complete multigrid cycle or SSOR methods. This is argued through a heuristic argument and illustrated by numerical results.
0 references
Helmholtz equation
0 references
variable index of refraction
0 references
finite element method
0 references
radiation boundary condition
0 references
preconditioned conjugate gradient method
0 references
iterative method
0 references
preconditioner
0 references
discrete Laplacian
0 references
multigrid method
0 references
relaxation scheme
0 references
complete multigrid cycle
0 references
0 references