In this paper a novel iterative method of multilevel type for solving large-scale generalized eigenvalue problems encountered in structural dynamics is presented. A preconditioned iterative technique, which can be viewed as a modification of the Subspace Iteration method, is used for simultaneous ca
A REDUCTION METHOD FOR LARGE SCALE UNSYMMETRIC EIGENVALUE PROBLEMS IN STRUCTURAL DYNAMICS
โ Scribed by Z.C. Zheng; G.X. Ren; W.J. Wang
- Publisher
- Elsevier Science
- Year
- 1997
- Tongue
- English
- Weight
- 354 KB
- Volume
- 199
- Category
- Article
- ISSN
- 0022-460X
No coin nor oath required. For personal study only.
โฆ Synopsis
The discussion begins with the classification of eigenvalue problems arising from conservative and non-conservative structural systems. The conservative type includes undamped structural eigenvalue problems and undamped gyroscopic eigenvalue problems. The non-conservative type includes damped structural eigenvalue problems, damped gyroscopic eigenvalue problems and constrainedly damped eigenvalue problems. The methods for solving large scale unsymmetric eigenvalue problems are briefly reviewed. The advantages and properties of Arnoldi's method have also been discussed. Arnoldi's reduction method has been generalized and the partial solution of large scale unsymmetric-definite eigenvalue problems in structural dynamics is presented in detail. A very simple reduction algorithm is obtained by simplifying the proposed method for undamped gyroscopic eigenvalue problems. To make the proposed reduction method feasible for engineering problems, a restart technique is introduced to work with Arnoldi's reduction method for checking and computing missing eigenvalues. Numerical examples are also presented to demonstrate the effectiveness of the proposed reduction method.
๐ SIMILAR VOLUMES
A new simple explicit two-step method and a new family of predictor-corrector integration algorithms are developed for use in the solution of numerical responses of dynamic problems. The proposed integration methods avoid solving simultaneous linear algebraic equations in each time step, which is va
A perturbation method to reduce the computing time for the eigenvalue analysis of structural dynamics is introduced in this paper. A simplified structural model, in which the degrees of freedom associated with the large stiffnesses are eliminated from the eigenvalue equation and the small masses are