Design Of Dynamic Fuzzy Neural Network Speed Controller For PMLSM

Linear motor servo equipments possess obvious advantages such as simple structure, low noise, fast speed, high precision, various combinations, etc. Therefore, linear motor servo system is applied widely. The uncertainties inside linear motor and outside disturbances are the main factors to reduce the system performances. In order to enhance the servo abilities of system and make it possesses excellent functional modes, the uncertainties and disturbances must be suppressed effectively. Dynamic fuzzy neural network (D-FNN), through its own parameters learning, changeable topological structure and etc, is able to guarantee the demands for performances under the severe situations arising from the uncertainties inside linear motor and outside disturbances. In this paper, the methods of D-FNN designed are applied to the linear motor servo system and further researched.At first, based on reading many domestic and international relative literatures the improvements of linear servo system are overviewed and the influences of main disturbance factors of linear motor servo control system to the dynamic qualities are analyzed; the domestic and international research status of permanent magnet linear synchronous motor (PMLSM) servo system based on D-FNN are outlined; the control schemes of PMLSM servo system are described. Secondly, on the basis of analyzing the basic structure and work principle of PMLSM, the vector control principle and the mathematic model of PMLSM are introduced; Thirdly, the theoretical foundations of D-FNN speed controller design are introduced, which include fuzzy logic and fuzzy inference; the structure and learning rules of neural network. Based on these theories, the method of D-FNN for PMLSM servo system is researched and D-FNN speed controller is designed according to the characters of nonlinear mathematical model for PMLSM and complicated status for disturbances. At last, the simulation analysis of D-FNN for PMLSM servo system is made with the help of Simulink software in Matlab. The simulation results show that D-FNN applied in PMLSM servo system is able to effectively suppress various disturbances, possess strong robustness and satisfy precise processing requirements for plants.