This paper describes a variable-damping dynamic absorber applying Electrorh
eological (ER) fluid to the damping element of a conventional-type dynamic
absorber. A prototype of variable-damping dynamic absorber was constructed
and its performance was verified with a three-story structural model. The s
pecial ability of the present dynamic absorber is to reduce the vibration a
mplitude at several frequencies by a single dynamic absorber. ER fluid is f
unctional fluid whose yield shear stress can be changed by the applied elec
tric field strength. Because of its peculiar property, ER fluid has been ap
plied to various mechanical components such as shock absorbers and engine m
ounts for vehicles, clutches, valves, etc. One of the practical ways in app
lying ER fluid to mechanical components may be to expand the performance of
conventional mechanical components by combining ER fluid effectively with
them. In this sense, this paper shows a successful application of ER fluid
to a conventional-type dynamic absorber. An adaptive neural network control
system composed of a forward model network for system identification and a
controller network was introduced to control the variable-damping element
of the dynamic absorber. The numerical simulations show good agreements wit
h the experimental results.