Affiliation:
1. Harbin Institute of Technology
2. RIGOL
Abstract
In order to overcome the system nonlinear instability and uncertainty inherent in magnetic bearing systems, two PID neural network controllers (BP-based and GA-based) are designed and trained to emulate the operation of a complete system. Through the theoretical deduction and simulation results, the principles for the parameters choice of two neural network controllers are given. The feasibility of using the neural network to control nonlinear magnetic bearing systems with un-known dynamics is demonstrated. The robust performance and reinforcement learning capability in controlling magnetic bearing systems are compared between two PID neural network controllers.
Publisher
Trans Tech Publications, Ltd.
Reference9 articles.
1. SCHWEITZER G, BLEULER H, TRAXLER A. Active magnetic bearings basics, properties and application of active magnetic bearings[M]. ETH, Switzerland：Hochschulverlag AG, (1994).
2. HAWKINS L A, MURPHY B T, KAJS J. Analysis and testing of a magnetic bearing energy storage flywheel with gain-scheduled MIMO control[C]/ Proceeding of ASME Turbo Expo, Munich, Germany：IGTI, 2000, 5：1-8.
3. SU Yixin, LONG Xiang, ZHANG Danhong, et al. Neural network adaptive PID control of active magnetic bearings[J]. Journal of Central China Normal University (Nat. Sci. ), 2004, 38(3): 304-307. (in Chinese).
4. A. Escalante. V. Guzman. M. Parada et al. Neural network emulation of a magnetically suspended rotor. Transactions of the ASME: Journal of Engineering for Gas Turbines and Power, 2004, 126: 373~384.
5. J. W. Kim, D. J. Xuan, Y. B. Kim. Control of a magnetic flywheel by a fuzzy neural network algorithm. ICMIT 2005: Control Systems and Robotics, Proc. of SPIE 2005, 6042: 29~45.