Parallel implementation of Galerkin technique in large-scale electromagnetic problems

Kaklamani, DI; Nikita, KS; Marsh, A

dc.contributor.author	Kaklamani, DI	en
dc.contributor.author	Nikita, KS	en
dc.contributor.author	Marsh, A	en
dc.date.accessioned	2014-03-01T01:48:41Z
dc.date.available	2014-03-01T01:48:41Z
dc.date.issued	1999	en
dc.identifier.issn	10569170	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/25561
dc.relation.uri	http://www.scopus.com/inward/record.url?eid=2-s2.0-0033225551&partnerID=40&md5=653c0afed8717c9b1b11dae95a3e131c	en
dc.subject	Code Parallelisation	en
dc.subject	Electrically Large Structures	en
dc.subject	Method of Moments	en
dc.subject.other	Dielectric materials	en
dc.subject.other	Frequency domain analysis	en
dc.subject.other	Functions	en
dc.subject.other	Galerkin methods	en
dc.subject.other	Integral equations	en
dc.subject.other	Linear systems	en
dc.subject.other	Matrix algebra	en
dc.subject.other	Method of moments	en
dc.subject.other	Natural frequencies	en
dc.subject.other	Parallel processing systems	en
dc.subject.other	Problem solving	en
dc.subject.other	Code parallelisation	en
dc.subject.other	Electrically large structures	en
dc.subject.other	Electromagnetic field theory	en
dc.title	Parallel implementation of Galerkin technique in large-scale electromagnetic problems	en
heal.type	journalArticle	en
heal.publicationDate	1999	en
heal.abstract	An integral equation formulation in conjunction with a parallelised Galerkin technique is employed to solve large-scale electromagnetic (EM) problems. The proposed technique is applicable to EM structures consisting of similar conducting or dielectric parts, defined as ""elements"". Coupled integral equations are derived in the frequency domain, written in terms of the conductivity currents or the electric fields developed on the conducting or dielectric ""elements"" surfaces, respectively. The system of integral equations is numerically solved via the parallel computed Galerkin technique, with convenient entire domain basis functions. Even for electrically large structures, the use of entire domain basis functions leads to relatively small order linear systems and the main computational cost refers to the matrix fill rather than the matrix solution. The parallelisation introduced to the computation of the matrix elements overcomes the limitation of using Method of Moments at lower and resonant frequencies. The inherent parallelism of the introduced technique allows for the results to be obtained with minimal additional to the sequential code programming effort. Two indicative electromagnetic compatibility applications are presented, concerning the coupling of incident waves with multiple conducting rectangular plates and the coupling phenomena occurring in a multi-element waveguide array looking into a layered lossy cylinder. Numerical results are presented, while the applicability/suitability of diverse High Performance Computing platforms is judged, based on both performance obtained and ease of code portation.	en
heal.journalName	Applied Computational Electromagnetics Society Newsletter	en
dc.identifier.volume	14	en
dc.identifier.issue	3	en
dc.identifier.spage	108	en
dc.identifier.epage	116	en