The Neural Network Of PMSM Vector Control System Based On MRAS

In recent years, High-performance permanent magnet synchronous motor(PMSM) control system has become a mainstream development direction of the electrical drive control system. Permanent magnet synchronous motor is a nonlinear, multivariable, strong coupling, time-varying plant, the controlled system is vulnerabled to factors of uncertainty motor parameter variations and external load disturbance, Traditional speed control control performance is not satisfactory, Therefore, the study of high-performance permanent magnet synchronous motor control system address the impact of these uncertain factors become a hot topic in the field of motor control. This thesis focuses on studying the high performance of PMSM vector control system, in order to improve the performance of the control system, and make the control system has strong anti-interference ability, high reliability and robustness.In this thesis, the space vector pulse width modulation(SVPWM) and permanent magnet synchronous motor mathematical model are considered as a theoretical basis, and bonding the theory of neural network and the theory of model reference adaptive(MRAS) to research a speed control scheme of permanent magnet synchronous motor based on neural network model reference adaptive control. The scheme based on error back propagation mechanism of BP neural network, discussing the BP neural network model reference adaptive speed identification method and taking the single neuron as speed controller, with emphasis on the control system of neural network identification algorithms to improve, since a BP neural network has the properties of forward propagation of message and backwark propagation fo errors, the parameters of single neuron can be modified to get proper values. The BP identifier is also used as a channel for error propagation. The algorithm uses the neural network which can approximate any complex nonlinear control object, and by self-learning online continuously, it can overcome the shortcomings of traditional model reference adaptive control algorithm, enabling the syste m to have a stronger robustness and the fault-tolerance.Then establishing simulation model in Matlab/simulink, the simulation results show that compared with the traditional MRAS speed identification method and the traditional PI speed control strategy, the proposed scheme has good identification precision and better adaptive ability.Finally, set up the hardware experimental platform with TMS320F2812 as the core chip, designing the hardware system as well as software system respectively, verifying the conclusion with related experiments. The experimental results show that the control strategy in this thesis has better adaptability, reliability, anti-jamming capability and robustness.