Variable-step Neural Network Control for Multi-frequency Transmission Forces of Multi-span Rotor Systems

WANG Jinjian,XU Hui,ZHU Changsheng
DOI: https://doi.org/10.3901/jme.2021.04.010
2021-01-01
Journal of Mechanical Engineering
Abstract::When the rotor system is in operation, the vibration of the base will be caused by the rotor's multi-frequency excitation forces which transmitted to the base through support bearings.A variable-step neural network algorithm based on tracking filtered journal displacements and transmission forces error signals is proposed to suppress the multi-frequency transmission forces.Firstly, a hybrid bearing integrating electromagnetic actuator and journal bearing is designed and its dynamic model is established.Secondly, the dynamic model of a multi-span rotors system is built by the finite element method, and the theoretical principle of reducing the transmission forces is analyzed.The discrete tracking differentiator and the state observer are used to solve the problem that the displacement signal is not easy to differentiate with white noise and reconstruction of the journal displacement at the actuator position when the sensor and actuator are non-collocated.Then a variable-step neural network algorithm for suppressing the multi-frequency transmission forces from the multi-span rotor system is proposed.Finally, numerical simulations are carried out in a four-bearing double-span rotor system to verify the effectiveness of variable-step neural network algorithm for suppressing the multi-frequency transmission forces from the rotor system to the base.The simulation results show that the variable-step neural network algorithm can effectively suppress the multi-frequency transmission forces from the rotor system to the base. Key words:multi-span
What problem does this paper attempt to address?