Magnetorheological (MR) and electrorheological (ER) dampers are known to exhibit nonlinear behaviour which can make it difficult to predict their performance, particularly when they are integrated into engineering structures. As a result it can be impossible to properly assess the feasibility of using such semi-active devices to solve practical engineering problems. In this paper, a new model format is proposed which represents an extension of earlier work by the authors. The proposed model is more general and yet maintains the physical significance of key parameters. A novel model updating (or system identification) technique is developed so that the model can account for the behaviour of various configurations of device without the need for prior knowledge of the fluid properties. The technique relies upon the iterative adjustment of the model’s stiffness parameter so that the quasi-steady behaviour of the device can be estimated. Correlation between a bi-viscous model and the estimated quasi-steady behaviour is used as the criterion for choosing the most suitable value of stiffness. The modelling technique is completed by establishing empirical shape relationships between the pre-yield parameters, post-yield parameters, yield force and the applied excitation conditions. The modelling and identification procedures are applied to an MR damping device and the results are validated by comparing predicted and experimental responses under both non-sinusoidal and broadband excitation conditions.

No abstract