Papers by Author: Ze Bin Yang

Abstract: A novel decoupling control method based on neural networks inverse system is presented in this paper for a bearingless synchronous reluctance motor (BSRM) possessing the characteristics of multi-input-multi-output, nonlinearity, and strong coupling. The dynamic mathematical models are built, which are verified to be invertible. A controller based on neural network inverse is designed, which decouples the original nonlinear system to two linear position subsystems and an angular velocity subsystem. Furthermore, the linear control theory is applied to closed-loop synthesis to meet the desired performance. Simulation and experiment results show that the presented neural networks inverse control strategy can realize the dynamic decoupling of BSRM, and that the control system has fine dynamic and static performance.

Abstract: The bearingless synchronous reluctance motor (BSynRM) is a multi-variable, nonlinear and strong-coupled system. To solve the difficult problem of precise decoupling in electric torque and radial suspension force control, the theory of direct torque control of traditional synchronous reluctance motor are applied to the torque control of a BSynRM in this paper. Based on mathematical models of the BSynRM, A direct torque control algorithm based on space vector pulse width modulation (SVM-DTC) is deduced. The SVM-DTC control system is designed and simulated. The simulation results show the control algorithm of SVM-DTC realizes decoupling in electric torque and radial suspension force control. The static and dynamic performance of electric torque, speed and radial suspension force of the BSynRM is excellent.

Abstract: In this paper, a novel radial hybrid magnetic bearing is proposed, which is driven by an AC 3-phase power converter, and biased magnetic fluxes are provided by an axial magnetized permanent magnet ring. Based on the introduction of the configuration and principle of the radial AC hybrid magnetic bearing (AC HMB), the methods in parameter design and design results for the prototype are presented. The parameter design results for the prototype of the radial AC HMB are simulated by 3D electromagnetic finite element method and the simulation results show that the obtained parameter is reasonable. The manufactured prototype applying to the parameter is operated and the test results also show that the design result for the prototype is available, and the AC HMB has good dynamic and static performance.

Abstract: A dynamic decoupling control method based on neural network inverse system theory is developed for the 5 degrees of freedom (5-DOF) rotor system. The rotor system suspended by AC hybrid magnetic bearings (HMBs) is a multivariable, nonlinear and strong coupled system. Firstly, the configuration of 5-DOF HMBs and the mathematical equations of suspension forces are set up. Secondly, it is demonstrated the system is reversible by analyzing mathematical model. On the basis, the neural network inverse system which is composed of the static neural networks and integrators, and original system are in series to constitute pseudo linear systems. Finally, linear system theory is applied to these linearization subsystems for designing close-loop controllers. The simulation results show that this kind of control strategy can realize dynamic decoupling control, and control system obtains good dynamic and static performances.