MODELING OF MAGNETORHEOLOGICAL DAMPER USING MULTIPLE ARTIFICIAL NEURAL NETWORKS FOR FUZZY CONTROL DESIGN OF A SEMI-ACTIVE SUSPENSION SYSTEM

Abstract

Suspension is a very important part of the functionality of a car. It supports the comfort of the passengers and can be used to improve the car's stability. If a magnetorheological (MR) damper is used, the damping performance can be adjusted according to the road condition. In this research, an attempt to model an MR damper using multiple artificial neural networks (ANNs) was conducted. The measurement data was obtained by using a damper dynamometer. Eight measurements were conducted to cover the coil current range between 0 and 1,400 mA. The MR damper model was then integrated into the semi-active suspension of a quarter-car model. Furthermore, the fuzzy logic controller (FLC) design for the semi-active suspension system was conducted. This artificial intelligence control deployed a novel approach in using the frequency content of the road surface as the input. The control scheme was tested by giving a sinusoidal bumpy road and a pseudo-random bumpy road as the input to the quarter-car model. The FLC performed well and was able to reduce the transmitted excitation from the road to the car chassis.