PID Parameters Optimization on Magnetic Suspension Pump Based on the BP Neural Network Model

Yi-xin SU,Yan-hui MA,Qian SHI,Shu XUE,Su-yuan YU
DOI: https://doi.org/10.3969/j.issn.1005-0329.2018.01.005
2018-01-01
Abstract:To design better performance dynamic of magnetic bearing support system of magnetic suspension pump, the multi-function magnetic levitation experiment platform was used to realize the 5-DOF stability suspension of the rotating shaft by setting the PID control parameters of the magnetic bearing. Then,the Y-direction of the radial magnetic bearing B was randomly selected to identify the BP neural network model of the controlled object,and PID parameters were optimized by Simulink. Finally,the simulation and comparison under three different working conditions were carried out. The simulation results show that the response of the system with optimized parameters is faster. And the experimental results show that the rotor's displacement in X and Y direction are reduced by 32.9% and 24.1% respectively in the course of 0~4000 r/min. This article provides a novel method for optimizing the PID parameters of the magnetic suspension pump.
What problem does this paper attempt to address?