Study On AC Servo Driven System And It's Novel Control Strategy

AC motor is a high order, nonlinear, coupling system, so the control strategy is most important in AC servo systems. In order to realize better characteristic of the system, the higher performance control strategy is expected. Combined with fuzzy logic, neural networks and sliding-mode control, some control methods are designed. The content of this research falls into the following:At first, the mathematical model of AC motor is presented which is based on the study of the physical and torque equations.A robust radial basis function neural network control system is developed for the robust and precise position control of the induction motor. In the proposed radial basis function control system, first, the ideal feedback linearization control law is designed based on the backstepping technique. Second, a radial basis function neural network controller is designed as the main controller, which is used to substitute the ideal feedback linearization control law. The effectiveness of the proposed control strategy is demonstrated by simulated results.To achieve robust control performance for a field-oriented AC motor servo drive system, a sliding-mode controller based on immune genetic algorithms is developed. First, an adaptive sliding-mode controller based IGA is proposed. An adaptive algorithm is utilized to estimate the bound of uncertainties, and a real-time IGA is developed to search the favorable adaptive parameter and switching parameter. The effectiveness of the proposed control strategy is demonstrated by simulated results.The fuzzy logic controller based neural network can make up for the disadvantages in the fuzzy logic by modifying the membership function in the fuzzy logic controller, because the neural network has the ability to be trained. The fuzzy neural network controller based on IGA is developed by on-line tuning of the fuzzy neural network learning rates and weights using IGA. The effectiveness of the proposed control strategy is demonstrated by simulated and experimental results.