Application of hierarchical matrices in the boundary element method for sound scattering from underwater objects

Translated title
Anwendung hierarchischer Matrizen in der Randelementmethode zur Untersuchung der Schallstreuung von Unterwasserobjekten
Publication date
2026-08-26
Document type
Konferenzbeitrag
Author
Rahlf, Paul
Christophersen, Sven
Börm, Steffen
Schäfer, Ingo
Ehrlich, Jan
Organisational unit
DOI
URI
Conference
189th Meeting of the Acoustical Society of America - joint with the Acoustical Society of Japan ; Honolulu, Hawaii ; December 1–5, 2025
Project
Publisher
Acoustical Society of America (ASA)
Series or journal
Proceedings of Meetings on Acoustics
Periodical volume
60
Article ID
022007
Peer-reviewed
✅
Part of the university bibliography
✅
Language
English
DDC Class
534 Schall und verwandte Schwingungen
Keyword
Boundary element method
Hierarchical Matrices
Underwater Acoustics
Target echo strength
Abstract
In order to determine the monostatic Target Echo Strength (TES) of underwater objects, the backscattering behavior of the structure(s) must be determined for many evaluation points, whereby these also define the respective positions of the sound source. The Boundary Element Method (BEM) is an established procedure for simulating this acoustic problem, but leads to dense complex matrices whose order corresponds to the number of surface elements. Each desired evaluation point (i.e. each sound source position) leads to a corresponding “right-hand side” for the system of equations. Depending on frequency and size, the number of required elements quickly becomes so large that the main memory of available workstations is insufficient to store the matrix and the solutions when using direct linear solution methods. In order to avoid this problem, the use of so-called hierarchical matrices was investigated as an alternative. Due to specialized compression algorithms, these matrices have significantly lower memory requirements and, when combined with iterative solvers (such as GMRES), should lead to faster solutions.This paper presents the results obtained from sample structures and compares the different solution methods in terms of memory usage, solution quality and computational performance.
Description
This is an open access article licensed under a Creative Commons Attribution 4.0 International (CC BY) License (https://creativecommons.org/licenses/by/4.0/).
Version
Published version
Access right on openHSU
Metadata only access