Fuzzy modeling of a magnetorheological damper using ANFIS

The magnetorheological (MR) damper is a semi-active control device that has received much attention by the vibration control community. Of primary interest is its fast response to a variable control signal as well as its low power requirements. The highly nonlinear dynamic nature of this device, however, has proven to be a significant challenge for researchers who wish to characterize its behavior. Research by others has shown that a system of nonlinear differential equations can successfully be used to describe the behavior of a MR damper. The paper presents an alternative for modeling a damper in the form of a Takagi-Sugeno-Kang fuzzy inference system. An ANFIS (adaptive neuro-fuzzy inference system) is used to determine 27 nonlinear premise parameters and 96 linear consequent parameters that describe the behavior of the SD-1000 model MR damper. Data used for training and checking of the model is generated from numerical simulation of nonlinear differential equations. The resulting fuzzy inference system is shown to satisfactorily represent behavior of the magnetorheological damper while greatly reducing computational requirements. Use of the neuro-fuzzy model increases the feasibility of real time simulation.