Optimization of a Fuzzy PID Controller

Dongxu Liu,Kai Zhang,Jinping Dong
DOI: https://doi.org/10.1007/978-3-642-21765-4_122
2011-01-01
Abstract:There are strong nonlinear characteristics in an active magnetic bearing (AMB) system. 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). Generally a fuzzy PID controller can perform better than a PID controller during the initial levitation process. But it’s difficult to design and optimize a fuzzy PID controller, because it mostly depends on experts’ knowledge and experiences. For a one DOF levitation ball model, the membership functions (MFs) were modified, the fuzzy rules were optimized, and some other ways were also tried to improve the performances of the fuzzy PID controller. And it turned out that the most effective one is to optimize the fuzzy rules. The modified fuzzy rules enhanced the anti-shock abilities of the fuzzy PID controller greatly, and at the same time the controller still performs well during the initial levitation process.
What problem does this paper attempt to address?