F. Brancaleoni

An Interval Uncertainty Based Method for Damage Identification

The problem of damage identification in presence of uncertainties is faced up in the framework of interval analysis. A method previously developed by the authors in the context of model updating and global minimization for dynamic problems is applied to identify the damage in framed structures. The inclusion property of the interval analysis is exploited to find the bounds of the physical solutions. Model parameters, experimental measures and modelling errors are considered as possible sources of uncertainty. The advantages of the interval approach are discussed through numerical simulations involving the different kind of uncertainties.

Model updating of Pescara Benchmark: interval vs. traditional method

In this work FE model updating is used to detect parameter variations from representative models of the Pescara benchmark beams. For a considered series of the beams experimental modal data are available at a certain beam stage. For this stage model updating results are obtained and different updating techniques are compared. Standard techniques based on parameters sensitivity calculation are applied and, in this case, commercial robust software and self-implemented are used. A new interval model updating approach is also applied to possibly cope with experimental uncertainties and parameters variability. The interval algorithm automatically gives a bound of the results; on the contrary sensitivity approach gives a crisp result. Hence, to have a proper comparison, a mid term comparison is also given from the calculated interval solution.