Neuro-fuzzy Active Control of Rotor Suspended on Active Magnetic Bearing

Abstract

This study deals with the active control of structures by using the capacities of artificial intelligence, and draws from the considerable potential of combining neural networks with fuzzy logic. In particular it focuses on the neuro-fuzzy controller design method and is part of a wider study aimed at emphasizing its feasibility for controlling the flexible modes of rotating machines at rest and in operation. The study covers the design, adjustment and application of this type of nonlinear controller that can be adjusted automatically. Here, the neuro-fuzzy control is used to control the operating position of a flexible rotor suspended on active magnetic bearings. The results of the simulation, confirmed by those of experiments, highlight both efficiency and performance of this type of controller and particular emphasis is given to the simplicity of its design.