Research On Neural Network Control System Of Hybrid Stepping Motor

The stepping motor servo system is widely used in industry. However, this system has some flaws, such as low frequency resonance which closes to the operating frequency, poor stationarity during low-speed running, bad rapid response capacity in high-speed running and less ability to resist loads’ fluctuations etc, so that its application is restricted. Researchers have been looking for more effective ways of control to solve these problems. In this thesis, some of new control method will be introduced.First of all, the two-phase hybrid stepping motor mathematical model is established referring to Leenhouts model and its parameters are determined by measurement and product manual. Then PID control strategy based on BP neural network is proposed. BP network can train the PID parameters. It’s an effective ways of online tuning PID parameters.Then, according to the motor mathematical model and BP-PID control method, the simulation system is accomplished on MATLAB/Simulink. The simulation results show this method can better reduce motor torque and speed fluctuation and the system has a good performance of tracking and anti-jamming abilities.Finally, the system platform has been built, which includes some hardware such as the motor drive and controller. And the software is PID control strategy integrating the BP algorithm. The experiments’ results also show this control mode can improve the capacity of anti-load impact, dynamic performance, and the running performance of the hybrid stepping motor.