WYD method for an eigen solution of coupled problems

Designing efficient and stable algorithm for finding the eigenvalues andeigenvectors is very important from the static as well as the dynamic aspectin coupled problems. Modal analysis requires first few significant eigenvectorsand eigenvalues while direct integration requires the highest value toasc...

Full description

Bibliographic Details
Main Authors: A Harapin, J Radnic, D Brzovic
Format: Article
Language:English
Published: Multi-Science Publishing 2016-04-01
Series:International Journal of Multiphysics
Online Access:http://journal.multiphysics.org/index.php/IJM/article/view/95
id doaj-100cf670dda941a38e6ff216ae296dfc
record_format Article
spelling doaj-100cf670dda941a38e6ff216ae296dfc2020-11-24T22:35:52ZengMulti-Science PublishingInternational Journal of Multiphysics1750-95482048-39612016-04-013210.1260/175095409788837801114WYD method for an eigen solution of coupled problemsA Harapin0J Radnic1D Brzovic2University of Split, Faculty of Civil Engineering and Architecture, Matice hrvatske 15, 21000 Split, CroatiaUniversity of Split, Faculty of Civil Engineering and Architecture, Matice hrvatske 15, 21000 Split, CroatiaUniversity of Split, Faculty of Civil Engineering and Architecture, Matice hrvatske 15, 21000 Split, CroatiaDesigning efficient and stable algorithm for finding the eigenvalues andeigenvectors is very important from the static as well as the dynamic aspectin coupled problems. Modal analysis requires first few significant eigenvectorsand eigenvalues while direct integration requires the highest value toascertain the length of the time step that satisfies the stability condition.The paper first presents the modification of the well known WYDmethod for a solution of single field problems: an efficient and numericallystable algorithm for computing eigenvalues and the correspondingeigenvectors. The modification is based on the special choice of thestarting vector. The starting vector is the static solution of displacements forthe applied load, defined as the product of the mass matrix and the unitdisplacement vector. The starting vector is very close to the theoreticalsolution, which is important in cases of small subspaces.Additionally, the paper briefly presents the adopted formulation for solvingthe fluid-structure coupled systems problems which is based on a separatesolution for each field. Individual fields (fluid and structure) are solvedindependently, taking in consideration the interaction information transferbetween them at every stage of the iterative solution process. The assessmentof eigenvalues and eigenvectors for multiple fields is also presented. This eigenproblem is more complicated than the one for the ordinary structural analysis,as the formulation produces non-symmetrical matrices.Finally, a numerical example for the eigen solution coupled fluidstructureproblem is presented to show the efficiency and the accuracy ofthe developed algorithm.http://journal.multiphysics.org/index.php/IJM/article/view/95
collection DOAJ
language English
format Article
sources DOAJ
author A Harapin
J Radnic
D Brzovic
spellingShingle A Harapin
J Radnic
D Brzovic
WYD method for an eigen solution of coupled problems
International Journal of Multiphysics
author_facet A Harapin
J Radnic
D Brzovic
author_sort A Harapin
title WYD method for an eigen solution of coupled problems
title_short WYD method for an eigen solution of coupled problems
title_full WYD method for an eigen solution of coupled problems
title_fullStr WYD method for an eigen solution of coupled problems
title_full_unstemmed WYD method for an eigen solution of coupled problems
title_sort wyd method for an eigen solution of coupled problems
publisher Multi-Science Publishing
series International Journal of Multiphysics
issn 1750-9548
2048-3961
publishDate 2016-04-01
description Designing efficient and stable algorithm for finding the eigenvalues andeigenvectors is very important from the static as well as the dynamic aspectin coupled problems. Modal analysis requires first few significant eigenvectorsand eigenvalues while direct integration requires the highest value toascertain the length of the time step that satisfies the stability condition.The paper first presents the modification of the well known WYDmethod for a solution of single field problems: an efficient and numericallystable algorithm for computing eigenvalues and the correspondingeigenvectors. The modification is based on the special choice of thestarting vector. The starting vector is the static solution of displacements forthe applied load, defined as the product of the mass matrix and the unitdisplacement vector. The starting vector is very close to the theoreticalsolution, which is important in cases of small subspaces.Additionally, the paper briefly presents the adopted formulation for solvingthe fluid-structure coupled systems problems which is based on a separatesolution for each field. Individual fields (fluid and structure) are solvedindependently, taking in consideration the interaction information transferbetween them at every stage of the iterative solution process. The assessmentof eigenvalues and eigenvectors for multiple fields is also presented. This eigenproblem is more complicated than the one for the ordinary structural analysis,as the formulation produces non-symmetrical matrices.Finally, a numerical example for the eigen solution coupled fluidstructureproblem is presented to show the efficiency and the accuracy ofthe developed algorithm.
url http://journal.multiphysics.org/index.php/IJM/article/view/95
work_keys_str_mv AT aharapin wydmethodforaneigensolutionofcoupledproblems
AT jradnic wydmethodforaneigensolutionofcoupledproblems
AT dbrzovic wydmethodforaneigensolutionofcoupledproblems
_version_ 1725722650576158720