ABSTRACT
Adaptive Neuro-Fuzzy Inference System (ANFIS) is used to design fuzzy reduced order
observer for solving the problem of unbalance in magnetic bearing systems. Nonlinear
model of magnetic bearing with unbalance is very complicated and coupling between
various states cannot be neglected. With classical methods no one can eliminate
disturbances without linearization of the system and neglecting some terms from the
nonlinear model. In this paper ANFIS is used to acquire desired membership functions
of the fuzzy rules for reduced order observer to achieve rejection of unbalance.
Simulation results show that fuzzy observer can eliminate the effect of unbalance at any
rotational speed.