Uncertainty Analysis in Structural Dynamics

Abstract

This paper discusses the main issues of Uncertainty Analysis (UA) in general and also argues and illustrates its particular relevance to structural dynamics. Brief descriptions are given of the most prevalent of the many frameworks for uncertainty representation. The three main uncertainty-related problems of relevance to structural dynamics are then discussed, namelyquantification,fusionandpropagation. In order to illustrate the application of ideas of UA in a realistic scenario, there then follows a case study conducted on an aerospace structure, namely the wing of a Gnat trainer aircraft. The case study considers evidence-based classifiers as an alternative to probabilistic classifiers for the problem of damage location within the context of Structural Health Monitoring. Dempster-Shafer theory is employed to construct neural network classifiers with the potential to admit ignorance, rather than misclassify.