Condensation methods for the dynamic analysis of large models
Calculations of the dynamic behaviour of a superstructure including soil-structure interaction with locally discretized models, e.g. using finite elements, lead to large-sized matrices. The order of the matrices may be so high, especially when three-dimensional modelling of soil is required, that ra...
Gespeichert in:
Veröffentlicht in: | Nuclear engineering and design 1989-02, Vol.111 (2), p.265-271 |
---|---|
Hauptverfasser: | , , |
Format: | Artikel |
Sprache: | eng |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | Calculations of the dynamic behaviour of a superstructure including soil-structure interaction with locally discretized models, e.g. using finite elements, lead to large-sized matrices. The order of the matrices may be so high, especially when three-dimensional modelling of soil is required, that rapid loss of accuracy occurs and exceptionally large computer storage and time are needed for calculations. To overcome these problems the order of the models can be reduced to the wanted degrees-of-freedom, for example to those of the superstructure. Condensation methods combined with subsystem modelling lead to models which avoid the calculation disadvantages mentioned above. Some of these methods are discussed and one of them is given preference.
This method is based on modal synthesis and reduces the magnitude of the truncation error by using the static parts of the truncated modes without the necessity of their being known. |
---|---|
ISSN: | 0029-5493 1872-759X |
DOI: | 10.1016/0029-5493(89)90213-6 |