Time series analysis of electrorheological actuator with double driving discs

Abstract

An electrorheological(ER) actuator with double driving discs rotating at the same speed in the opposite directions was designed for studying the electrorheological torque experiments. The dynamic model of the disc transmission system consists of the linear part between the output angle and the electric torque, and the nonlinear part between the electric torque and the applied field strength dependent on the ER fluids. Using only the output sampled torque signals, an autoregressive model, the auto-spectral density function, the autoregressive(AR) bispectrum and the slices of the bispectra, taken to analyse the torque dynamic response, are obtained when a zero mean and non-Gaussian white noise interferes with the rotary disc system. The method for AR model order selection based on bispectral cross correlation is proposed and employed to determinate the model order. The experimental and theoretical results show that the time series analysis method and the parametric bispectrum might be helpful to establish the dynamic model of an ER actuator and to quantitatively analyse the torque response.