- Title
- A comparative study of preconditioning techniques for large sparse systems arising in finite element limit analysis
- Creator
- Kardani, Omid; Lyamin, Andrei V.; Krabbenhøft, Kristian
- Relation
- IAENG International Journal of Applied Mathematics Vol. 43, Issue 4, p. 195-203
- Relation
- http://www.iaeng.org/IJAM/issues_v43/issue_4/index.html
- Publisher
- Newswood
- Resource Type
- journal article
- Date
- 2013
- Description
- The efficiency of several preconditioned Conjugate Gradient (PCG) schemes for solving of large sparse linear systems arising from application of interior point methods to nonlinear Finite Element Limit Analysis (FELA) is studied. Direct solvers fail to solve these linear systems in large sizes, such as large 2D and 3D problems, due to their high storage and computational cost. This motivates using iterative methods. However, iterative solvers are not efficient for difficult problems without preconditioning techniques. In this paper, the effect of various preconditioning techniques on the convergence behavior of the preconditioned Conjugate Gradient (PCG) is investigated through a detailed comparative study. Furthermore, numerical results of applying PCG to several sample systems are presented and discussed thoroughly in a parametric study. Our results suggest that while incomplete Cholesky preconditioners are by far the most efficient techniques for sequential computations, significant gains may result from use of sparse approximate inverse methods in parallel environment in this field.
- Subject
- incomplete Cholesky factorization; approximate inverse preconditioner; limit analysis; preconditioned conjugate gradient method; cone programming
- Identifier
- http://hdl.handle.net/1959.13/1061059
- Identifier
- uon:16869
- Identifier
- ISSN:1992-9978
- Language
- eng
- Reviewed
- Hits: 2832
- Visitors: 3063
- Downloads: 0
| Thumbnail | File | Description | Size | Format |
|---|