Abstract:
This paper presents an approach to approximate the inverse dynamic behaviour of a magneto-rheological
(MR) damper using radial basis function (RBF) networks: Due to the highly nonlinear characteristics of
MR dampers, modelling of MR dampers becomes a very important problem to their applications. In this
paper, an alternative representation of the MR damper in te,nns of RBF networks, which have a structure
of four input neurons and one output neuron to emulate the inverse dynamic behaviours of an MR damper
is developed. Training and validating of the RBF network models are achieved by using the data
generated from the numerical simulation of the nonlinear differential equations proposed for the MR
damper. The centres of the RBF networks are searched by using different methods such as subtractive
clustering algorithm, k-means clustering algorithm and fuzzy c-means clustering algorithm. The sum of
squared errors (SSEs) between the true outputs and the network predictions are compared for different
network structures and different approaches. It is shown by the validation tests that the RBF networks can
represent the inverse dynamic behaviours of the MR damper satisfactorily.
