A Fuzzy Controller Design For Active Magnetic Bearing System

Liu Dongxu,Zhang Kai,Dong Jinping
2010-01-01
Abstract:There are strong nonlinear characteristics in an active magnetic bearing (AMB) system. Even for a simple magnetic levitation ball model with one degree of freedom (DOF), there are problems that are difficult to be solved by using linear control methods (such as PID control method). When linear control methods are applied, small overshoots and static errors are not easily obtained at the same time especially in the initial levitation process, when control currents of the linear controller have to be reasonable. In order to find the reasons and improve the controller's performance, a single pole and a dual-pole one DOF magnetic levitation ball system were used for simulations. The simulation results based on linear and nonlinear magnetic force model were provided and analyzed from the following aspects in an AIM system: control signals of a PID controller, magnetic forces of the magnet and kinetic energy of the ball. On this basis, a fuzzy PID controller for the dual-pole with nonlinear magnetic force model system was designed using fuzz logic combined with a ND controller. The fuzzy inference system could be used to tune PID parameters according to the system's states. When the ball's displacements or velocities were different, different PID parameters were applied. By comparing the simulation results of a fuzzy PID controller and a PID controller, it turned out that a fuzzy PID controller could solve the problem of the ND controller during the initial levitation process and achieved better performance in the one DOF magnetic levitation ball system.
What problem does this paper attempt to address?