Loading…
A Fourth-Order Kernel-Free Boundary Integral Method for the Modified Helmholtz Equation
Based on the kernel-free boundary integral method proposed by Ying and Henriquez (J Comput Phys 227(2):1046–1074, 2007 ), which is a second-order accurate method for general elliptic partial differential equations, this work develops it to be a fourth-order accurate version for the modified Helmholt...
Saved in:
Published in: | Journal of scientific computing 2019-03, Vol.78 (3), p.1632-1658 |
---|---|
Main Authors: | , |
Format: | Article |
Language: | English |
Subjects: | |
Citations: | Items that this one cites Items that cite this one |
Online Access: | Get full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | Based on the kernel-free boundary integral method proposed by Ying and Henriquez (J Comput Phys 227(2):1046–1074,
2007
), which is a second-order accurate method for general elliptic partial differential equations, this work develops it to be a fourth-order accurate version for the modified Helmholtz equation. The updated method is in line with the original one. Unlike the traditional boundary integral method, it does not need to know any analytical expression of the fundamental solution or Green’s function in evaluation of boundary or volume integrals. Boundary value problems under consideration are reformulated into Fredholm boundary integral equations of the second kind, whose corresponding discrete forms are solved with the simplest Krylov subspace iterative method, the Richardson iteration. During each iteration, a Cartesian grid based nine-point compact difference scheme is used to discretize the simple interface problem whose solution is the boundary or volume integral in the BIEs. The resulting linear system is solved by a fast Fourier transform based solver, whose computational work is roughly proportional to the number of grid nodes in the Cartesian grid used. As the discrete boundary integral equations are well-conditioned, the iteration converges within an essentially fixed number of steps, independent of the mesh parameter. Numerical results are presented to verify the solution accuracy and demonstrate the algorithm efficiency. |
---|---|
ISSN: | 0885-7474 1573-7691 |
DOI: | 10.1007/s10915-018-0821-8 |