A Procedure to Assess Total Uncertainty in Characterizing Structural Response

This paper summarizes an on-going development of a metric to quantify total uncertainty in the validation of a computer code that might be used, for example, to predict the amount of structural damage in a constructed facility undergoing some levels of loading. The total uncertainty is expressed as a combination of various forms of uncertainty. Such a combination is based on a group of methods collectively called generalized information theories. Among these theories are evidence theory, possibility theory, fuzzy set theory and, of course, probability theory. A simple case study on the crushing of a thin foam material is used to illustrate how the validation of a computational prediction can be assessed by the representation and quantification of uncertainties that influence the difference between the prediction and the experimental results. In the case study, damage is expressed in terms of excessive accelerations in the foam