Neural Network Modeling of a Magnetorheological Damper

This paper presents the results of modeling a magnetorheological (MR) damper by means of a neural network. TheMR damper is a device extensively used to mitigate hazardous vibrations in systems such as vehicles, bridges and buildings. The advantages of these dampers over other devices of their class rely on the cost, size, simplicity and performance. However, these devices are highly nonlinear; their dynamics is characterized by friction and hysteresis, phenomena that has been difficult to model following physical laws. In this paper, a recurrent neural network is trained to reproduce the behavior of the damper and it will be shown that the results are better that those obtained by approximate physical models.

[1]  Shirley J. Dyke,et al.  Modeling and identification of the hysteretic dynamics of an MR actuator for its application to semiactive control of flexible structures , 2007, SPIE Smart Structures and Materials + Nondestructive Evaluation and Health Monitoring.

[2]  Shirley J. Dyke,et al.  Experimental Identification of a Shear-Mode MR Damper and Numerical Evaluation of Hysteretic Models , 2007 .

[3]  James Lam,et al.  Modelling of a magneto-rheological damper by evolving radial basis function networks , 2006, Eng. Appl. Artif. Intell..

[4]  Shirley J. Dyke,et al.  Phenomenological Model of a Magnetorheological Damper , 1996 .

[5]  P. N. Roschke,et al.  Fuzzy modeling of a magnetorheological damper using ANFIS , 2000, Ninth IEEE International Conference on Fuzzy Systems. FUZZ- IEEE 2000 (Cat. No.00CH37063).

[6]  Wei-Hsin Liao,et al.  Neural network modeling and controllers for magnetorheological fluid dampers , 2001, 10th IEEE International Conference on Fuzzy Systems. (Cat. No.01CH37297).

[7]  Billie F. Spencer,et al.  Models for hysteresis and application to structural control , 1997, Proceedings of the 1997 American Control Conference (Cat. No.97CH36041).