Study On Commutation Torque Ripple Minimization Techniques For Permanent Magnet Brushless DC Motor

Permanent magnet brushless DC motor has the advantages of large torque density, low noise, high efficiency and so on. So it is widely used in the numerical control machine tools, instrumentation, aerospace, household appliances and other aspects. But the torque ripple problem during the operation of the permanent magnet brushless DC motor has great influence to achieve accurate location and high performance speed control. This article gives a research about inhibiting permanent magnet brushless dc motor commutation torque ripple problem.Firstly, this paper has a detailed analysis of the permanent magnet brushless DC motor’s basic structure, working principle and the drive method, including the establishment of the permanent magnet brushless DC motor mathematical model, and provides a theoretical foundation for the new control strategy.Secondly, the permanent magnet brushless DC motor commutation torque ripple of causes are analyzed in detail in this paper, and knowing that maintain the commutation current constant is the key to suppress commutation torque ripple.This paper puts forward a intelligent control method of the neural network current prediction to suppress the commutation torque ripple, this method has a function of the compensation adjustment for three phases current,maintain the commutation phase current constant during commutation, and then suppress the commutation torque ripple.Finally, this paper designs a dual-loop control system of permanent magnet brushless DC motor for the torque ripple suppression,the outer ring of the speed loop, using PID control based on BP Neural Network; inner current loop, using combination of PI and commutation current predictive control strategy. It completes a comparative simulation of traditional PID control and neural network commutation current predictive control while permanent magnet brushless DC motor running at high speed and low-speed areas. The simulation results show that the neural network commutation current predictive control is suitable for any speed range of permanent magnet brushless DC motor, it is a good suppression of the motor commutation torque ripple and improves the system stability and anti-jamming capability. At the same time, the thesis completes the current and speed control experiments based on DSP platform.