Study Of Fuzzy Control Method Based On Neural-network For Permanent Magnet Synchronous Motor

Permanent magnet synchronous motor (PMSM) has been widely used in high performance drive applications for its advantages, such as compactness, high efficiency, reliability and suitability to environment. In order to achieve high performance, the vector control of PMSM drive is employed. However, the controller design of the system plays crucial role in the system performance. Some machine parameters (for example rotor resistance) and drag load parameters (for example moment of inertia) often change with the different work conditions. At the same time, PMSM machine is a nonlinear controlled object. It is very difficult for traditional, linear and fixed gain PID controller to achieve the satisfactory control effect of higher precision. But the intelligent controller can improve system adaptation to some disturbance and nonlinear factor, because it can self-adaptively change parameters and cover the deficiencies of linear fixed gain PID. As the important embranchment of intelligent control, fuzzy control based on neural network will have wide development future, because which synthesizes the merit of fuzzy control and neural network. So the study of fuzzy control method based on neural-network to PMSM has important theory and practical meaning.The paper presents a fuzzy control method based on neural-network to PMSM. First, the basic theory of coordinate conversion, mathematical model of PMSM and basic theory of vector control is introduced. The i_d = 0 vector control scheme based on rotor field orientationis designed. The simulation model of current and speed double loops PMSM control system is made up in Matlab/Simulink. Second, the fuzzy self-adjusting PID controller is designed based on fuzzy PID theory as the speed controller of current and speed double loops PMSM vector control system. The simulation model of system is built in MATLAB/SIMULINK. The feasibility of fuzzy self-adjusting PID control method is proved. Then, neural network PID controller is designed and the PMSM is considered as a second-order transfer function. The system simulation is run using the m files of MATLAB. At last, the fuzzy controller based onneural-network is designed based on the study of fuzzy self-adjusting PID controller and neural network PID controller. The fuzzy controller based on neural-network remember fuzzy control rulers using neural network, which synthesizes the merits of fuzzy logic and neural network control. It has the characteristic of controlled object mathematical model-independence, learning ability and adaptability to the external environment. The fuzzy controller based on neural-network is used as the speed controller of PMSM current and speed double loops vector control system and the simulation is run and analyzed. The simulation results show the fuzzy control based on neural-network of PMSM has better response speed and overshoot than fuzzy control, which can better overcome external disturbance and parameter variation and has good dynamic and static performance.