Vibration control of a structure with an MR damper using a neural network

Magnetorheological ( MR) fluids are known as one of highly functional materials of which the yield shear stress can be varied in accordance with applied magnetic field strength. Because of their peculiar properties, MR fluids have been applied to various kinds of mechanical devices such as brakes, clutches, valves, etc. In this paper, a prototype of a variable damper using an MR fluid has been constructed and its typical performance has been measured. A three-story structural model is equipped with the MR damper, and the structural response has been measured under various conditions of excitation frequency, amplitude and damping properties produced by the MR damper. Furthermore, a neural 1 network has been applied to an adaptive vibration control system for the structure and the performance of the control system has been verified by experiments.