Nonparametric modeling of magnetorheological damper based on nonlinear black-box technique

Abstract

Abstract Magnetorheological (MR) dampers are considered the most confident actuators for controlling vibratory systems due to their great advantages like high controllable dynamic range, low power consumption, and fail-safe devices. In this paper, a new nonparametric technique is utilised to model the MR-damper dynamics using Non-Linear Autoregressive Models with Exogenous Inputs (NLARX) approach. The proposed model is constructed based on the measured displacement, velocity, and damping forces of the commercial MR damper under various command voltages. The results of the nonparametric model show that such a model can expect the nonlinear dynamics of the MR damper under a wide range of different operating conditions. Furthermore, the proposed model output is compared with a well-known parametric model namely as Modified Bouc-Wen (MBW) model for additional validation. The comparison shows that the generated model can accurately track the measured damping force characteristics, work done by the damper, and the force-velocity hysteresis. It is also shown that the improvement of the damping force of the proposed model is better than the corresponding damping force of the parametric model which clearly shows the robustness of the generated model. The proposed model improves the root mean square (RMS) values of the experimental damping force by about 93.57% compared to an improvement of about 77.36% of the simulated damping force from the MBW model.