Abstract
There is a pressing need to improve the feasibility of three-dimensional finite element (FE) methods applied to many problems in civil engineering. This is particularly the case for static analyses in geotechnical engineering: ideally, models would be 3D, follow the actual geometry, use non-linear material
formulations and allow simulation of construction sequences, and all of this with a reasonable degree of accuracy. One major obstacle to improvements in this regard is the difficulty in solving of the set of (linearised) algebraic equations which arises from a typical discretisation approach. Very large systems
become cumbersome for direct techniques to solve economically. This paper describes the incorporation of iterative (rather than direct) solution techniques, developed through University research, into commercial FE software for geotechnics.
formulations and allow simulation of construction sequences, and all of this with a reasonable degree of accuracy. One major obstacle to improvements in this regard is the difficulty in solving of the set of (linearised) algebraic equations which arises from a typical discretisation approach. Very large systems
become cumbersome for direct techniques to solve economically. This paper describes the incorporation of iterative (rather than direct) solution techniques, developed through University research, into commercial FE software for geotechnics.
Original language | English |
---|---|
Title of host publication | Proceedings of 17th UK conference on computational mechanics in engineering |
Editors | C. Sansour |
Number of pages | 4 |
Publication status | Published - 2009 |
Event | 17th UK Conference on Computational Mechanics, ACME UK - London, UK Duration: 6 Apr 2009 → 8 Apr 2009 |
Conference
Conference | 17th UK Conference on Computational Mechanics, ACME UK |
---|---|
City | London, UK |
Period | 6/04/09 → 8/04/09 |
Keywords
- iterative solvers
- commercial software
- preconditioning
- geomechanics
- commercial fe code