System Of Neural Defuzzification Maglev Spherical Motor

With the development of industrial automation, the manipulator which is driven by the traditional single degree of freedom motor has not meet the requirements, now it puts forward higher requirements in manufacturing. In this case, the magnetic levitation motor appeared, compared with traditional ordinary motor, there is no friction in the motor which uses the magnetic levitation technique when it is running, so it doesn’t need any lubricating oil which increases the work life, this motor can be used in some high speed and ultra-high speed occasions; and because the motor uses the spherical structure which can perform multiple degrees of freedom movement, so it can avoid the complex machines which is caused by the traditional single degree of freedom motor and simplifies the structure of whole system. Because of the above characteristics, it has become a current research hot spot.This paper is based on the structure of the magnetic levitation spherical switched motor, because the control system of the magnetic levitation motor is a complex system with the nonlinear and high coupling, routine control has not yet reached technological requirements. So this paper mainly introduces the fuzzy neural inverse system, but different from the traditional fuzzy neural network, in the condition of the invertibility, the paper first design the model of neural network inverse system to control the levitation force and rotational, utilizing the function approximating model of artificial neural network and the decoupling linearization of the inverse system, and on this basis, fuzzy PID is designed, in order to achieve the optimal control effect.This paper’s main work mainly includes the following three parts; the first paper is to use the principle of the air-gap magnetic field to solve the levitation force and torque after gaining insight into the structure and operating principle of the magnetic levitation spherical switched motor, and according to the characteristics of magnetic suspension spherical motor with single winding, differential control strategy is put forward and on this basis, the inverse is introduced, so this paper uses the neural network to make inverse system decoupling, and fuzzy PID is designed; according the requirement of the control system the second paper design the hardware and software design of the digital signal processor based on the main control chip; the last part make the simulation of the control system in MATLAB/Simulink module and verify the superiority of the control system, and provide guidance for practical application.