Interval fields to represent uncertainty on input and output side of a FE analysis

Intervals have been used extensively for the representation of parametric uncertainty in Finite Element Models when the available information is insufficient to build representative probabilistic models. Although the complexity of the interval concept is rather limited, the numerical problem arising...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Verhaeghe, Wim, Rousounelos, Andreas, Desmet, Wim, Vandepitte, Dirk, Moens, David
Format: Tagungsbericht
Sprache:eng
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Intervals have been used extensively for the representation of parametric uncertainty in Finite Element Models when the available information is insufficient to build representative probabilistic models. Although the complexity of the interval concept is rather limited, the numerical problem arising at the core of the Interval Finite Element Method was proven to be non-trivial. Many research activities have focussed on this problem over the last decade. Still, one of the principal shortcomings of the current interval finite element procedures is that they are intrinsically not capable of representing dependencies among uncertain input and/or output quantities. This poses important limitations on the applicability of the interval concept in non-deterministic finite element analysis. These limits are both surfacing in the model definition phase as well as in the post-processing phase. This paper introduces the concept of interval fields for the static analysis of uncertain mechanical structures in the context of the Finite Element Method. The theoretical background of the concept is presented, and it is shown how implicit and explicit interval fields can be used to represent dependent uncertainties in the model definition phase and in the post-processing phase, thus allowing for the calculation of sharper bounds on derived results. Furthermore, the applicability and accuracy of both types of interval fields are studied by means of a numerical validation example.