Control System Research Of Bldcm Based On Artificial Neural Network

Brushless DC Motor is a new type motor which is developed on base of DC motor with the advancement of the power electronics, motor control technology and microprocessor. To brushless DC motor, the prediction of moter's properties is affected by some parameters, such as back electromotive force, etc.When brushless DC motor and its control system are simulated in traditional methods, air-gap induction usually is hypothesized as sine-wave or square-wave distribution. As a matter of fact, the air-gap induction of brushless DC motor is nonlinear distribution. Therefore, there is a big error between this hypothesis and the real situation. The traditional speed regulator of brushless DC Motor adopts PID controller, but during the motor running, the parameter of the motor and disturbance always constantly vary. So it's difficult to get the satisfactory control effect by the traditional speed PI regulator.Aiming at the problem above, the simulation model of brushless DC motor is constructed firstly based on analyzing the mathematic model of brushless DC motor by MATLAB/SIMUL INK in this paper. The speed and electric current double closed-circle is adopted to control the brushless DC motor in the system.Secondly, making use of neural network's the property of approaching nonlinear function at random precision, the back electromotive force of brushless DC motor is predicted by combining BP neural network with genetic algorithm in this paper. The predicted results indicated that the back electromotive force of brushless DC motor is more precise in this method.Finally, to enhance the performance of brushless DC motor drives, traditional PID control is combined organically with the neural network in this paper, which structured the single neural cell PID controller in order to cope with shortcomings of traditional PID control. Simulation results indicate that the improved PID control mentioned above is better than traditional one.