Comparison of Magnetorheological Damper Models through Parametric Uncertainty Analysis Using Generalized Likelihood Uncertainty Estimation

Abstract

Magnetorheological (MR) dampers provide a viable alternative for vibration control of structures subjected to external excitations. Realistic modeling of MR dampers is critical for control development and structural design. Many existing phenomenological models of MR dampers are optimized from experimental observations for deterministic parameter values. However, vibration control design based on deterministic values might not provide reliable response prediction due to inherent parameter uncertainties from optimization. Parametric uncertainty analysis of damper models using laboratory experiments enables quantification of potential effect on vibration control design and evaluation. Several existing MR damper models were evaluated for their parametric uncertainties based on characterization tests of a 200-kN large-scale MR damper. Markov-chain Monte Carlo (MCMC) simulation was integrated with the generalized likelihood uncertainty estimation (GLUE) method to derive the posterior distribution of model parameters. Prediction of MR damper behavior was used to illustrate the influence of model parametric uncertainties. The MR damper models were compared in terms of energy dissipation in the real-time tests with predefined displacements.