2003
DOI: 10.1002/nme.804
|View full text |Cite
|
Sign up to set email alerts
|

Improved conditioning of infinite elements for exterior acoustics

Abstract: SUMMARYAn optimized version of the so-called mapped wave envelope elements, also known as Astley-Leis elements, is introduced. These elements extend to inÿnity in one dimension and therefore provide an approach to the simulation of exterior acoustical problems in both frequency and time domains. Their formulation is improved signiÿcantly through the proper choice of polynomial bases in the direction of radiation. In particular, certain Jacobi polynomials are identiÿed which behave well with respect to conditio… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
28
0

Year Published

2007
2007
2017
2017

Publication Types

Select...
5
1
1

Relationship

0
7

Authors

Journals

citations
Cited by 49 publications
(28 citation statements)
references
References 42 publications
0
28
0
Order By: Relevance
“…The condition number κ(A) is given by the ratio of the largest to the lowest eigenvalue of the frequency-dependent matrix A and is calculated by power iteration and inverse iteration, respectively. As shown by Babuška et al 41 Cremers et al 42 and Dreyer et al 6 the matrix condition increases with the polynomial order, which leads to ill-conditioned matrices for polynomials of degree > 8 in the case of Lagrange polynomials, whereas Legendre and Jacobi polynomials provide distinctly better-conditioned matrices. Tests performed by the authors of this paper have confirmed this observation, although the choice between the four polynomials had no considerable effect on the modes in exterior acoustics, as will be shown subsequently.…”
Section: Matrix Condition Numbermentioning
confidence: 98%
See 2 more Smart Citations
“…The condition number κ(A) is given by the ratio of the largest to the lowest eigenvalue of the frequency-dependent matrix A and is calculated by power iteration and inverse iteration, respectively. As shown by Babuška et al 41 Cremers et al 42 and Dreyer et al 6 the matrix condition increases with the polynomial order, which leads to ill-conditioned matrices for polynomials of degree > 8 in the case of Lagrange polynomials, whereas Legendre and Jacobi polynomials provide distinctly better-conditioned matrices. Tests performed by the authors of this paper have confirmed this observation, although the choice between the four polynomials had no considerable effect on the modes in exterior acoustics, as will be shown subsequently.…”
Section: Matrix Condition Numbermentioning
confidence: 98%
“…Investigations by von Estorff and Dreyer et al 6,34 prove poor suitability of Lagrangian polynomials for higher orders and achieve better performance with Legendre and Jacobi polynomials. This was shown by calculating the condition number of the dynamic stiffness matrix A in the system of linear equations in Eq.…”
Section: Acoustics and Discretization Of The Unbounded Domainmentioning
confidence: 99%
See 1 more Smart Citation
“…Various choices of Q j (x) have been investigated, including Lagrangian, 78,79 Legendre, 82 Jacobi, 83 and rational (integrated Jacobi). 84 Lagrangian shape functions result in very poorly conditioned infinite element matrices.…”
Section: D71 Infinite Element Shape Functionsmentioning
confidence: 99%
“…The other three choices all appear to give acceptable levels of conditioning. Dreyer 83 showed that the Jacobi polynomials in general give a better condition than the Legendre polynomials. Regardless of the choice for Q(x), equations D.57 and D.64 imply that P(x) will be a function of the master element coordinates r, s,t, and thus can be integrated over the master element.…”
Section: D71 Infinite Element Shape Functionsmentioning
confidence: 99%