Fuzzy Model Updating: covariance updating to estimate the interval radii of the updated parameters

摘要

Stochastic model updating techniques aim at reducing the epistemic uncertainty mainly related to the lack of knowledge. These techniques are usually computationally demanding and the use of non-probabilistic methods allows for the reduction of the computational effort, typically associated with stochastic model updating methods. This paper presents a fuzzy model updating method based on a convex joint fuzzy membership of all the experimental responses and on the sensitivity-based updating of the interval centre at each alpha-cut of the uncertain model parameters. Moreover, the covariance updating concept is used to estimate the interval radii of the referred intervals directly from the covariance matrix of the experimental response set. The developed method allows to achieve precise predictions of the intervals of the parameters with a reduced computational effort. A numerical example is given to evaluate the implementation and performance of the proposed fuzzy model updating method.