The Research On Radial Basis Function Neural Network AC Servo Speed Controller System

With the development of industry and astronomy, performance and function of servo system are being concerned more and more. AC servo systems are used more widely for its excellent performance. AC servo system is a nonlinear system with significant coupling. To improve the quality of AC servo system, control strategy is very important. According to the characteristic of AC servo system, This paper combines Radial Basis Function Neural Network with vector control, and designs a self adaptive speed controller.Firstly, the mathematical model of PMSM is studied. Then, the paper analyses vector control principle of rotor field orientation and deduces servo control method of PMSM.Secondly, according to the characteristics of PMSM, RBF Neural Network is chose for its perfectly approaching to nonlinear function. The principles and learning algorithms are introduced in detail.Finally, on the basis of theoretical analysis, a self adaptive speed controller is designed and applied to a single chip AC servo control system. According to FPGA module design characteristics, the system is divided into some functional modules. The way of SVPWM and self adaptive speed controller are discussed which are key on servo system. Furthermore, the method of on-line and off-line training is proposed. At last, the single chip AC servo control system is applied to NC machine. The experiment results show that the control system has better performance and powerful robustness, which improves the servo system's dynamic property.