The Application Of The BP Neural Network In Brushless DC Motor Speed Control System

With the advantages of high torque, high precision, high efficiency etc., The BrushlessDC motor has played an important role in national defense, aerospace, robotics, precisionmachine tools and other fields. But the BLDCM itself is a nonlinear system, its parameterschange when working conditions change. It is difficult to describe in accurate systemmodel when magnetic circuit is saturation, load is disturbed etc. PID control method issimple and has good robustness and high reliability etc. characteristics. So it is wide usedin the field of industrial control. But the control actions of proportion, integral anddifferential cooperate and restrict each other. They are not simple linear combination, andcan be established under the condition of knowing the precise system mode. As amulti-variable, strong coupling, nonlinear system, It is hard to determine the PIDparameters of the BLDCM from nearly endless variety of nonlinear relationships. So it issignificant to develop the strategies of BLDCM speed control system.BP neural network has the advantage of approximating arbitrary nonlinear functions,which does not require accurate model of plant. It is the one which get the most commonused and achieve the best practical effect. It’s not easy to solve the practical problems ofthe BLDCM control system only relying on a kind of intelligent control method. So thispaper combine the BP neural network with the traditional PID control theory to realize theimportant parameters, Proportional, integral, differential constant self-learning andself-adjusting.The BP neural network PID controller is simulated in MATLAB on the theoreticalbasis. The simulation results reveal that the BP neural network PID speed controllersystem’s response speed is faster and fluctuation is smaller than traditional PID speedcontroller system. Then, a BLDCM speed control system is designed based onTMS320F28335DSP, and do experiments by the method this paper proposed.