Stochastic and hybrid methods for the identification in the Bouc-Wen model for magneto - Rheological dampers

Abstract

Magneto-rheological (MR) dampers are semi-active control devices that require low power input and are considered fail-safe, i.e. in case of control hardware failure they become passive dampers. In the present work we consider a sensitivity analysis and use the deterministic gradient based Levenberg-Marquardt method, the stochastic methods Simulated Arnealing and Genetic Algorithms, as well as hybridizations of these methods, for the estimation of parameters in the Bouc-Wen model, which is used to describe the dynamic behavior of MR dampers. The parameters associated with the histeretic displacement evolution equation are the most difficult to be estimated, and therefore we use a strategy for estimating such parameters separately from other parameters of the model. © 2008 IOP Publishing Ltd.